Contemporary Debates in Metaphysics Contemporary Debates in Philosophy
In teaching and research, philosophy makes progress through argumentation and debate. Contemporary Debates in Philosophy provides a forum for students and their teachers to follow and participate in the debates that animate philosophy today in the western world. Each volume presents pairs of opposing viewpoints on contested themes and topics in the central subfi elds of philosophy. Each volume is edited and introduced by an expert in the fi eld, and also includes an index, bibliography, and suggestions for further reading. The opposing essays, commissioned especially for the volumes in the series, are thorough but accessible presentations of opposing points of view.
1. Contemporary Debates in Philosophy of Religion edited by Michael L. Peterson and Raymond J. Vanarragon 2. Contemporary Debates in Philosophy of Science edited by Christopher Hitchcock 3. Contemporary Debates in Epistemology edited by Matthias Steup and Ernest Sosa 4. Contemporary Debates in Applied Ethics edited by Andrew I. Cohen and Christopher Heath Wellman 5. Contemporary Debates in Aesthetics and the Philosophy of Art edited by Matthew Kieran 6. Contemporary Debates in Moral Theory edited by James Dreier 7. Contemporary Debates in Cognitive Science edited by Robert Stainton 8. Contemporary Debates in Philosophy of Mind edited by Brian McLaughlin and Jonathan Cohen 9. Contemporary Debates in Social Philosophy edited by Laurence Thomas 10. Contemporary Debates in Metaphysics edited by Theodore Sider, John Hawthorne, and Dean W. Zimmerman
Forthcoming Contemporary Debates are in:
Political Philosophy edited by Thomas Christiano and John Christman Philosophy of Biology edited by Francisco J. Ayala and Robert Arp Philosophy of Language edited by Ernie Lepore Contemporary Debates in Metaphysics
Edited by Theodore Sider, John Hawthorne, and Dean W. Zimmerman © 2008 by Blackwell Publishing Ltd BLACKWELL PUBLISHING 350 Main Street, Malden, MA 02148-5020, USA 9600 Garsington Road, Oxford OX4 2DQ, UK 550 Swanston Street, Carlton, Victoria 3053, Australia The right of Theodore Sider, John Hawthorne, and Dean W. Zimmerman to be identifi ed as the authors of the editorial material in this work has been asserted in accordance with the UK Copyright, Designs, and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs, and Patents Act 1988, without the prior permission of the publisher. Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks, or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. First published 2008 by Blackwell Publishing Ltd 1 2008 Library of Congress Cataloging-in-Publication Data Contemporary debates in metaphysics / edited by Theodore Sider, John Hawthorne, and Dean W. Zimmerman. p. cm. — (Contemporary debates in philosophy) Includes bibliographical references and index. ISBN 978-1-4051-1228-4 (hardcover : alk. paper) — ISBN 978-1-4051-1229-1 (pbk. : alk. paper) 1. Metaphysics. I. Sider, Theodore. II. Hawthorne, John (John P.) III. Zimmerman, Dean W. BD95.C66 2007 110—dc22 2007019836 A catalogue record for this title is available from the British Library. Set in 10 on 12.5 pt Rotis Serif by SNP Best-set Typesetter Ltd., Hong Kong Printed and bound in Singapore by Markono Print Media Pte Ltd. The publisher’s policy is to use permanent paper from mills that operate a sustainable forestry policy, and which has been manufactured from pulp processed using acid-free and elementary chlorine-free practices. Furthermore, the publisher ensures that the text paper and cover board used have met acceptable environmental accreditation standards. For further information on Blackwell Publishing, visit our website at www.blackwellpublishing.com Contents
Notes on Contributors vii
Introduction Theodore Sider 1
ABSTRACT ENTITIES 9 1.1 Abstract Entities Chris Swoyer 11 1.2 There Are No Abstract Objects Cian Dorr 32
CAUSATION AND LAWS OF NATURE 65 2.1 Nailed to Hume’s Cross? John W. Carroll 67 2.2 Causation and Laws of Nature: Reductionism Jonathan Schaffer 82
MODALITY AND POSSIBLE WORLDS 109 3.1 Concrete Possible Worlds Phillip Bricker 111 3.2 Ersatz Possible Worlds Joseph Melia 135
PERSONAL IDENTITY 153 4.1 People and Their Bodies Judith Jarvis Thomson 155 4.2 Persons, Bodies, and Human Beings Derek Parfi t 177
TIME 209 5.1 The Privileged Present: Defending an “A-Theory” of Time Dean Zimmerman 211 5.2 The Tenseless Theory of Time J. J. C. Smart 226 PERSISTENCE 239 6.1 Temporal Parts Theodore Sider 241 6.2 Three-Dimensionalism vs. Four-Dimensionalism John Hawthorne 263
FREE WILL 283 7.1 Incompatibilism Robert Kane 285 7.2 Compatibilism, Incompatibilism, and Impossibilism Kadri Vihvelin 303
MEREOLOGY 319 8.1 The Moon and Sixpence: A Defense of Mereological Universalism James Van Cleve 321 8.2 Restricted Composition Ned Markosian 341
METAONTOLOGY 365 9.1 Ontological Arguments: Interpretive Charity and Quantifi er Variance Eli Hirsch 367 9.2 The Picture of Reality as an Amorphous Lump Matti Eklund 382
Index 397
Contents vi Notes on Contributors
Phillip Bricker is Professor and Head of Philosophy at the University of Massachusetts, Amherst. His interests range broadly over metaphysics, philosophical logic, philosophy of science, and philosophy of mathematics.
John W. Carroll is Professor of Philosophy at NC State University in Raleigh, North Carolina. He works in the areas of metaphysics and the philosophy of science. His interests center on the topics of laws of nature, causation, explanation, and time travel. He is the author of Laws of Nature (Cambridge University Press, 1994) and such articles as “Ontology and the Laws of Nature” (Australasian Journal of Philoso- phy, 1987), “The Humean Tradition” (Philosophical Review, 1990), “Property-Level Causation?” (Philosophical Studies, 1991), and “The Two Dams and that Damned Paresis” (British Journal for the Philosophy of Science, 1999). He is the editor of Readings on Laws of Nature (Pittsburgh University Press, 2004).
Cian Dorr received his BA from University College Cork, and his PhD from the Uni- versity of Princeton, where he was a student of the late David Lewis. He is currently Assistant Professor of Philosophy at the University of Pittsburgh.
Matti Eklund is an Associate Professor of Philosophy at Cornell University. He has published articles in metaphysics, philosophy of language, and philosophy of logic.
John Hawthorne is Waynfl ete Professor of Metaphysical Philosophy at the University of Oxford. He is author of Metaphysical Essays (Clarendon Press, 2006), and has published widely in metaphysics, epistemology, philosophy of language, and Leibniz studies. Eli Hirsch is Professor of Philosophy at Brandeis University. He is the author of a number of works in metaphysics, including Dividing Reality (Oxford University Press, 1993).
Robert Kane is University Distinguished Teaching Professor of Philosophy at the University of Texas at Austin. He is author of The Signifi cance of Free Will (Oxford University Press, 1996), Through the Moral Maze (Paragon House, 1994), A Contem- porary Introduction to Free Will (Oxford University Press, 2005) and editor of The Oxford Handbook of Free Will (2002), among other works in the philosophy of mind and ethics.
Ned Markosian is a philosophy professor at Western Washington University. He grew up in Montclair, New Jersey, graduated from Oberlin College, and received a PhD from the University of Massachusetts. He has worked mainly on issues in the philoso- phy of time and the mereology of physical objects.
Joseph Melia is a Reader in Metaphysics at the University of Leeds. His main interests are in modality, ontology, and the philosophy of physics. He is currently working on a book on ontology.
Derek Parfi t was born in China in 1942 and received an undergraduate degree in Modern History at Oxford in 1964. Since 1967 he has been a Fellow of All Souls College, Oxford. He has often taught in the United States, and is now a regular Visit- ing Professor to the Departments of Philosophy of Rutgers, New York University, and Harvard. His fi rst book, Reasons and Persons, was published by Oxford University Press in 1984. A second book, Climbing the Mountain, is nearly completed, and will also be published by Oxford University Press. This book will be about reasons and rationality, Kant’s ethics, contractualism, and consequentialism.
Jonathan Schaffer is Professor of Philosophy at the Australian National University. He works mainly in metaphysics and epistemology. Further information about his work may be found on his website:
Theodore Sider is Professor of Philosophy at New York University. He has pub- lished articles in metaphysics and philosophy of language, is the author of Four- Dimensionalism (Oxford University Press, 2001), and is co-author (with Earl Conee) of Riddles of Existence: A Guided Tour of Metaphysics (Oxford University Press, 2005).
J. J. C. Smart is Emeritus Professor, Australian National University, and is now living in Melbourne. He is an honorary in the School of Philosophy and Bioethics at Monash University. His most recent publication is a paper “Metaphysical Illusions” which is pertinent to the chapter in the present volume.
Chris Swoyer is Professor of Philosophy and Affi liated Professor of Cognitive Psy- chology at the University of Oklahoma. He has published, and continues to work on, Notes on Contributors viii the philosophy of logic, metaphysics, philosophy of science, and history of modern philosophy (especially Leibniz).
Judith Jarvis Thomson is Professor of Philosophy at MIT. Her published work is on topics in moral theory, metaethics, and metaphysics.
James Van Cleve taught for many years at Brown University and is now Professor of Philosophy at the University of Southern California. He works in metaphysics, epistemology, and the history of modern philosophy.
Kadri Vihvelin is Associate Professor of Philosophy at the University of Southern California. Her publications include “The Dif” (Journal of Philosophy, 2005); “Freedom, Foreknowledge, and the Principle of Alternate Possibilities” (Canadian Journal of Philosophy, 2000); “What Time Travelers Cannot Do” (Philosophical Studies, 1996); “Causes, Effects, and Counterfactual Dependence” (Australasian Journal of Philoso- phy, 1995); and “Stop Me Before I Kill Again” (Philosophical Studies, 1994).
Dean W. Zimmerman is an Associate Professor in the Philosophy Department at Rutgers University. He is editor of Oxford Studies in Metaphysics and author of numer- ous articles in metaphysics and philosophy of religion.
Notes on Contributors ix
Introduction
Theodore Sider
There is something strange about metaphysics. Two strange things, really, although they are related. Metaphysics asks what the world is like.1 But the world is a big and varied place. How can one meaningfully ask what apples, planets, galaxies, tables, chairs, air conditioners, computers, works of art, cities, electrons, molecules, people, societies . . . are like? The question is hopelessly general and abstract! One would normally ask fi rst what apples are like, and then ask what planets and the rest are like separately. What meaningful questions are there about such a broad and hetero- geneous subject matter? Furthermore, you’d think that you’d need to ask a biologist what apples are like, an astronomer what planets are like, and so on. What can a philosopher contribute? Let’s have a look. Consider a certain apple. What is it like? Well, it’s red, and it’s round. But this information doesn’t come to us from philosophy. We need to observe the apple to learn its color and shape. Consider another thing, Mars. It has iron oxide on its surface, and it is 6.4185 × 1023 kg in mass. This information about Mars, again, isn’t something that philosophy can tell us about; we learn it from astronomers. So far, we have found no philosophical subject matter. But if we abstract from certain details, we fi nd things in common between our two examples; we fi nd a recurring pattern despite the diverse subject matters. Here are the facts we cited:
The apple is red Mars has iron oxide on its surface The apple is round Mars is 6.4185 × 1023 kg in mass
Notice that in each case, an object is said to have a feature. For example, in the fi rst case, the object is the apple, and the feature is being red. Philosophers call objects that have features particulars, and they call the features “had” by particulars proper- ties. Thus, we have:
The apple is red Mars has iron oxide on its surface · · · · · · particular property particular property
The apple is round Mars is 6.4185 × 1023 kg in mass · · · · · · particular property particular property
In fact, this pattern is quite general. Think of other facts:
Fact particular property This table is broken the table being broken Electron e is negatively charged electron e negative charge The stock market crashed the stock market crashing
The particular-property pattern keeps recurring. It appears that every fact about the world boils down to particulars having properties.2 So it would seem that the world contains two different sorts of entities: particulars and properties. We have already uncovered a general fact about the world. Just as a scientist establishes generaliza- tions about what the world is like in some limited sphere (for instance that charged particles repel one another or that the planets move in elliptical orbits), we have established a generalization – albeit a much broader and more abstract one – about the world. And we did it without detailed input from the sciences. Of course, since this is philosophy we are talking about, there is controversy at every turn. The statement that there are two different sorts of objects in the world, particulars and properties, can be challenged. Nominalists, for example, believe in particulars, but not in properties. According to a nominalist, there simply is no such thing as the property of being red. Put that baldly, the statement is misleading. It suggests that nominalists think that there is no such thing as a red object. But nominalists are not crazy. They agree that red objects exist; they just deny that redness exists. The nominalist’s position can be made clearer by thinking about the sentence ‘The apple is red’. The nominalist agrees that the sentence is true. But now, consider the two parts of the sentence: its subject, ‘The apple’, and its predicate, ‘is red’. What the nominalist thinks is that, whereas the subject does stand for an object (namely, the particular in question, the apple), the predicate does not stand for an object. The predicate ‘is red’ is of course meaningful; it’s just that it doesn’t stand for an object. Just as a comma is meaningful without standing for an object, predicates can be meaningful without standing for objects. The apple is red, even though there is no such thing as its redness. We talk as if there are lots of things, when really, those things don’t exist. We talk, for instance, as if there are such things as holes. We’ll say: “Look at the size of that hole in the wall!” “Bring me the piece of cheese with three holes in it.” “I can’t wear that shirt because there is a hole in it.” But surely there aren’t really such things Theodore Sider 2 as holes, are there? What kind of object would a hole be? Surely what really exist are the physical objects that the holes are “in”: walls, pieces of cheese, shirts, and so on. When one of these physical objects has an appropriate shape – namely, a perfo- rated shape – we’ll sometimes say that “there is a hole in it.” But we don’t really mean by this that there literally exists an extra entity, a hole, which is somehow made up of nothingness. The nominalist thinks that all subject-predicate sentences are a bit like sentences about holes. It might seem at fi rst that the predicates refer to entities, but they really don’t. Are nominalists right? Do properties exist or don’t they? This is no easy question, and Chris Swoyer and Cian Dorr (chapter 1) come to opposite conclusions on this and related matters. But in this brief look at nominalism, we have at least glimpsed what metaphysicians are after: patterns in apparently diverse phenomena, and generaliza- tions that accurately describe these patterns. This book contains chapters in a number of areas of metaphysics; in each area, the goal is to fi nd generalizations about abstract patterns: Necessity Scientists tell us of the laws of nature. Physicists tell us of the laws of physics, for example that like-charged particles must repel one another. Chemists tell us of the laws of chemistry, for example that if methane reacts with oxygen, it must produce carbon dioxide and water. Economists tell us of the laws of economics, for example that when demand increases then prices must increase as well. In each case, we have scientists telling us what must happen in certain conditions. What exactly are these laws of nature; what is the status of these “musts”? Laws of society exist because governing bodies have legislated them. But there is no governing body that has leg- islated the laws of nature. Physicists try to discover the laws of physics; they do not create them (chapter 2). And if everything happens as these laws of nature specify, human actions must conform to their dictates. How then can we have free will (chapter 7)? Further, there are other cases of “mustness”. Every bachelor must be male; every prime number other than two must be odd. In what does the mustness of these facts consist (chapter 3)? Time Objects of all sorts, the objects of physics, chemistry, biology, and other sciences, last over time. This raises many philosophical questions. What does it mean for the same object to exist over time? A person at age 50, for instance, is the same person as she was as a child, even though nearly all of the matter that made up her body as a child no longer is with her at age 50. What makes a person the same over time? And indeed, what is it for time to pass at all (chapters 4–6)? Ontology Different sciences describe different objects. Physics describes subatomic particles, biology describes organisms, and so on. But must we believe that the objects from each science really exist? Consider organisms, for example. Could we not stick with the physicist’s objects, and say that the only objects that really exist are subatomic particles? We could still agree that there are distinctively biological phenomena, even though there do not exist distinctively biological objects. For even if human organisms Introduction 3 (for example) do not exist, there are nevertheless certain systems of particles that exhibit biological behavior. These are the systems involving particles that one ordi- narily thinks of as being parts of a single biological organism. Thus, we have very general ontological questions (existence questions) about objects with parts (chapter 8). Other ontological questions include the question discussed above of whether prop- erties exist, the question of whether numbers exist, and even the “metaontological” question of what it means to investigate whether objects of a certain sort “really” exist (chapter 9). Within these and other areas of metaphysics, certain themes recur. For example, metaphysicians tend to fall into two camps: those who go around trying to reduce phenomena, and those who prefer instead to “leave the world as they found it.” Con- sider the law of nature saying that like-charged particles repel one another. Of one thing we can be sure: the existence of such a law guarantees a regularity: everywhere and at any time, every pair of like-charged particles will indeed repel each other. Jonathan Schaffer (chapter 2.2) is a member of the reductionist camp. He wants to say that, roughly, there is nothing more to this law beyond the regularity. The law reduces to the regularity. What the physicists discover is simply that it is universally true that every two charged particles in fact repel each other. John W. Carroll disagrees (chapter 2.1); he is from the anti-reductionist camp. According to him, reductionists like Schaffer leave out something crucial. They leave out the mustness, the necessity, of laws. It doesn’t just happen to be the case that charged particles repel one another. When you give two particles the same charge, they must repel each other. So there’s something more to a law than just the fact that objects everywhere act in accordance with the law; you need to add necessity to a regularity to get a law. Another example: time’s passage. We ordinarily think of time as something that “moves”. J. J. C. Smart (chapter 5.2) takes a reductionist approach to time’s passage. According to him, time is just another dimension like space. And like space, it is not really correct to describe time as moving. What we ordinarily think of as time’s passage just arises from the fact that at any given moment in time, we can only remember what has occurred in one direction through time (the direction we call the “past”). But objects in this direction are not “gone.” Just as objects that are spatially distant – for example, objects on Mars – are just as real as objects around here, so, objects that are temporally distant – for example, dinosaurs – are just as real as objects around now. Dean Zimmerman, on the other hand, resists this reduction (chapter 5.1). Our ordinary belief about the matter is correct: time has passed since the time of the dinosaurs, and the dinosaurs are now gone. And this does not just mean that they are far away in time, just as Mars is far away in space. The dinosaurs simply do not exist. A second (and related) recurring theme in metaphysics is the relationship between a scientifi c outlook and our ordinary beliefs. What science tells us doesn’t always fi t neatly with our ordinary beliefs about the world. In cases of confl ict, should we revise science so that it doesn’t confl ict with our ordinary beliefs? Should we revise the ordinary beliefs in light of science? Or is it a mistake to think that they confl icted in the fi rst place? Time’s passage again provides an example. The picture of time we get from physi- cists, especially from Einstein’s theories of relativity, is Smart’s picture of space-like Theodore Sider 4 time. But where, in this picture, is there room for our ordinary belief that time passes? According to Smart, our ordinary belief must be revised to fi t it into the scientifi c picture, whereas according to Zimmerman, it is the scientifi c picture that must be revised, or at least augmented. Or consider the problem of free will and determinism. Science tells us of a world governed by laws of nature. An electron has no choice about where to move; if another charged particle is in its vicinity, it cannot help but be repelled. The laws of nature must be obeyed. But on the face of it, this threatens our ordinary conception of our- selves as having free choices. We blame evildoers because we think that their choices were not inevitable; they freely chose to do wrong. Robert Kane (chapter 7.1) argues that these two pictures genuinely confl ict. If the laws of nature fully determined what each and every object in the world was going to do, then there would be no room for any human freedom. (Fortunately, there is reason to think that the laws of nature that scientists have actually discovered are not quite so restrictive.) Kadri Vihvelin, on the other hand, tries to fi t human freedom into the world of science, even a sci- entifi c world in which all human behavior is determined (chapter 7.2). But Vihvelin does not think that this calls for a revision of our ordinary beliefs about freedom. (In this way her position is unlike Smart’s.) According to Vihvelin, it was a mistake to think that the two world-pictures were in confl ict in the fi rst place. What should we trust when doing metaphysics: science or ordinary beliefs? The question leads some to extremes. At one end, we fi nd those who think that all meta- physics can do is report science. At the other end, we fi nd those who think that metaphysics should ignore science and listen only to ordinary beliefs. Each extreme is questionable. The fi rst extreme ignores the fact that science does not settle all metaphysical questions, and also the fact that scientists are infl uenced by their metaphysical pre- suppositions. We need a metaphysics that goes beyond reporting science in order to address the unsettled questions and evaluate the presuppositions. The second extreme subdivides. It includes those who think that science and ordi- nary beliefs can never confl ict, because they address “different worlds” (the “world of ordinary life” and the “world of science”). And it includes those dogmatists who think that ordinary beliefs can never seriously be doubted. The problem with each subdivision is that neither ordinary beliefs nor science is intended to be about a novel subject matter. Each is about the world. Ordinary folks, naturally, have beliefs about the world; but they hope to learn more about it through science. In addition to believ- ing that objects move in space over time, that actions take time, and that objects take up space, ordinary believers also expect science to tell us the underlying nature of space and time. Nor do scientists step into another world when they don their lab coats. The point of science is to understand how the world, the one world, the world in which ordinary folks live, works. A moderate view of the relation between science and ordinary beliefs seems in order: metaphysics must listen to, but is not exhausted by, science. This, however, leaves the exact nature of the relation wide open. Perhaps ordinary beliefs are epis- temic starting points – claims with which we are entitled to begin our inquiries, but which may later be revised, perhaps because they confl ict with science, perhaps because they confl ict with one another. Perhaps not all ordinary beliefs should be Introduction 5 taken equally seriously. We might, for example, grant more weight to beliefs that are fundamental to the structure of our thought about the world (recall the discussion of particulars and properties above), and grant little (if any) weight to ordinary beliefs about matters more properly addressed by the sciences. Perhaps the mere fact that a belief is an ordinary one counts for nothing at all; perhaps we should instead trust reason, a faculty capable of guiding both philosophically sophisticated scientists and scientifi cally informed philosophers. Any metaphysician is bound, sooner or later, to face the following challenge. Science has been wildly successful. It has led to increasingly successful theories, technological advances, and consensus as to the truth. The history of metaphysics, on the other hand, has been as much one of wild goose chases as progress. Metaphy- sicians (like all philosophers!) continue to disagree about the same issues for millennia, and have not sent anyone to the moon. This leads some philosophers to doubt that metaphysics has any value at all. A certain empiricist tradition in epistemology says that the only route to truth is through the senses, and ultimately through science. If you can’t do an experiment to settle a question, the question isn’t worth asking. At best, it is an idle question whose answer we will never know; at worst, the question is meaningless. The empiricist is moved by an admirable desire to rid philosophy of undisciplined speculation. But the only empiricism that fl atly rules out all metaphysics is one based on a naive view of science. Real scientists do not just “summarize what they see.” Scientists must regularly choose between many theories that are consistent with the observed data. Their choices are governed by criteria like simplicity, comprehensive- ness, and elegance. This is especially true in very theoretical parts of science, for instance theoretical physics, not to mention mathematics and logic. A realistic picture of science leaves room for a metaphysics tempered by humility. Just like scientists, metaphysicians begin with observations, albeit quite mundane ones: there are objects, these objects have properties, they last over time, and so on. And just like scientists, metaphysicians go on to construct general theories based on these observations, even though the observations do not logically settle which theory is correct. In doing so, metaphysicians use standards for choosing theories that are like the standards used by scientists (simplicity, comprehensiveness, elegance, and so on). Emphasizing continuity with science helps to dispel radical pessimism about meta- physics; the humility comes in when we remember the discontinuities. Observation bears on metaphysics in a very indirect way, and it is far less clear how to employ standards of theory choice (like simplicity) in metaphysics than it is in science. But metaphysicians can, and should, acknowledge this. Metaphysics is speculative, and rarely if ever results in certainty. Who would have thought otherwise? Exactly what one should say about empiricism and metaphysics is a deep philo- sophical question in its own right, and it’s unlikely that anyone will decisively answer it anytime soon. But that shouldn’t, on its own, deter you from thinking about meta- physics. Philosophy is the one discipline in which questions about the value of that discipline are central questions within that very discipline. The philosopher must therefore live with uncertainty about whether her life’s work is ultimately meaningful – that is the cost of the breadth of refl ection demanded by philosophy. Philosophy’s Theodore Sider 6 refl ective nature is generally a good thing, but the down side is that it can lead to paralysis. Don’t let it. You don’t need to have answers to all meta-questions before you can ask fi rst-order questions ( just as you don’t need to sort out the philosophy of biology before doing good work in biology). The meta-questions are certainly important. But the history of philosophy is full of sweeping theories saying that this or that bit of philosophy is impossible. Take heart in the knowledge that these have all failed miserably.
Notes
1 As opposed to, for example, what the world ought to be like (ethics), what we know about the world (epistemology), how we think of and talk about the world (philosophy of mind and language), and so on. 2 Some facts consist of multiple particulars having a “multi-place” property, also known as a relation. Philadelphia is 100 miles from New York: the particulars Philadelphia and New York have the 100 miles from relation.
Introduction 7
CHAPTER ONE ABSTRACT ENTITIES
1.1 “Abstract Entities,” Chris Swoyer 1.2 “There Are No Abstract Objects,” Cian Dorr
“Concrete” entities are the entities with which we are most familiar: tables, chairs, planets, protons, people, animals, and so on. “Abstract” entities are less familiar: numbers (for example, the number seven), properties (for example, the property of being round), and propositions (for example, the proposition that snow is white). Do abstract entities really exist? No one has ever seen, touched, or heard an abstract entity; but Chris Swoyer argues that they exist nevertheless. Cian Dorr argues that they do not.
CHAPTER 1.1
Abstract Entities
Chris Swoyer
One of the most puzzling topics for newcomers to metaphysics is the debate about abstract entities, things like numbers (seven), sets (the set of even numbers), properties (triangularity), and so on. The major questions about abstract entities are whether there are any, if so which ones there are, and if any do exist, what they are like. My aim here is to provide a brief and accessible overview of the debates about abstract entities. I will try to explain what abstract entities are and to say why they are important, not only in contemporary metaphysics but also in other areas of phi- losophy. Like many signifi cant philosophical debates, those involving abstract entities are especially interesting, and diffi cult, because there are strong motivations for the views on each side. In the fi rst section, I discuss what abstract entities are and how they differ from concrete entities and in the second section, I consider the most compelling kinds of arguments for believing that abstract entities exist. In the third section, I consider two examples, focusing on numbers (which will be more familiar to newcomers than other types of abstract objects) and properties (to illustrate a less familiar sort of abstract entity). In the fi nal section, I examine the costs and benefi ts of philosophical accounts that employ abstract entities.1
1 What are Abstract Entities?
Prominent examples of abstract entities (also known as abstract objects) include numbers, sets, properties and relations, propositions, facts and states-of-affairs, pos- sible worlds, and merely-possible individuals (we’ll see what some of these are in a bit). Such entities are typically contrasted with concrete entities – things like trees, dogs, tables, the Earth, and Hoboken. I won’t discuss all of these examples, but will consider a few of the more accessible ones as case studies to help orient the reader. Numbers and sets Thought and talk about numbers are extremely familiar. We learn about the natural 2 numbers (like three, four, and four billion), about fractions (rational numbers, like /3 7 and /8), and about irrational numbers (like the square root of 2 and e). And we learned a bit about sets in school – for example, the empty set, the set containing just 3 and 4, and the set of even numbers; we even learned to write names of sets using notation like ‘{3,4}’. But what are numbers and sets? We cannot see them or point to them; they do not seem to have any location, nor do they interact with us or any of our instruments for detection or measurement in any discernible way. This may lead us to wonder whether there really are any such things as numbers, and whether, when we say things like “there is exactly one prime number between four and six,” we are literally and truly asserting that such a number exists (after all, what could it be?). But, as we will see in section 3.1, there are also strong philosophical arguments that numbers do exist. Hence a philosophical problem: do they or don’t they?
Properties and relations The world is full of resemblances, recurrences, repetitions, similarities. Tom and Ann are the same height. Tom is the same height now as John was a year ago. All electrons have a charge of 1.6022 × 10−19 coulomb. The examples are endless. There are also recurrences in relations and patterns and structures. Bob and Carol are married, and so are Ted and Alice; the identity relation is symmetrical, and so is that of similarity. Resemblance and similarity are also central features of our experience and thought; indeed not just classifi cations, but all the higher cognitive processes involve general concepts. Philosophers call these attributes of qualities or features of things (like their color and shape and electrical charge) properties. Properties are the ways things can be; similarly, relations are the ways things can be related. Assuming for the moment that there are properties and relations, it appears that many things have them. Physical objects: The table weighs six pounds, is brown, is a poor conductor of electricity, and is heavier than the chair. Events: World War I was bloody and was fought mainly in Europe. People: Wilbur is six feet tall, an accountant, irascible, and married to Jane. Numbers: three is odd, prime, and greater than two. All of these ways things can be and ways they can be related are repeat- able; two tables can have the same weight, two wars can both be bloody. The two adjacent diamonds in fi gure 1 are the same size, orientation, and uniform shade of gray. Champions of properties hold that things like grayness (or being gray) and trian- gularity (or being triangular) are properties, and that things like being adjacent and being a quarter of an inch apart are relations. Since the goal here is just to give one prominent example of a (putative) sort of abstract object, I will think of properties as universals (as many, but not all, philosophers do). On this construal, there is a single, universal entity, the property of being gray, that is possessed or exemplifi ed by each
ab
Figure 1 Resemblances and Ways Things Can Be Chris Swoyer 12 of the two diamonds in our fi gure. It is wholly present in both a and b, and will remain so as long as each remains gray. Philosophers who concur that properties exist may disagree about which properties there are and what they are like, but at least many properties (according to numerous philosophers, all) are abstract entities. Perhaps a property like redness is located in those things that are red, but where is justice, or the property of being a prime number, or the relation of life a century before? Such properties and relations exist outside space and time and the causal order, so they are rather mysterious. But, as we will see, there are also good reasons for thinking that properties and relations can do serious philosophical work, helping explain otherwise puzzling philosophical phenom- ena. This is a reason to think that they do exist. Another problem. Propositions Two people can use different words to say the same thing; indeed, they can even use different languages. When Tom says “Snow is white” and Hans says “Schnee ist weiss,” there is an obvious sense in which they say the same thing. So whatever this thing is, it seems to be independent of any particular language. Philosophers call these entities propositions. They are abstract objects that exist independently of language and even thought (though of course many of them are expressed in language). Propo- sitions have been said to be the basic things that are true or false, the basic truth- bearers, with the sentences or statements that express them being derivatively true or false. In addition to saying that snow is white, Tom also believes that snow is white; and Hans, who speaks no English, also believes that snow is white (although he expresses the belief by saying “Schnee ist weiss”). Again, there is an obvious sense in which they believe the same thing. Some philosophers urge that the best way to explain this is to conclude that there is some one thing that Tom and Hans both believe. On this view, propositions are said to be the contents or meanings of beliefs, desires, hopes, and the like. They are also said to be the objects of beliefs. Thus the object of Tom’s belief that red is a bright color is the proposition that red is a bright color. On this view propositions are abstract objects that express the meanings of sen- tences, serve as the bearers of truth values (truth and falsehood), and are the objects of belief. But like numbers, propositions are somewhat mysterious. We can’t see them, hear them, point to them. They don’t seem to do anything at all. This gives us reason to doubt their existence. But, there are also reasons to think that they exist. Problems, problems, problems.
1.1 What abstract entities are (nearly enough) Debates about abstract objects play a central role in contemporary metaphysics. There is wide agreement about the paradigm examples of abstract entities, though there is also disagreement about the exact way to characterize what counts as abstractness. Perhaps this shouldn’t come as a surprise; if any two things are so dissimilar that their difference is brute and primitive and hard to pin down, abstract entities and concrete entities (abstracta and concreta) are certainly plausible candidates. Even so, the philosophically important features of the paradigm examples of abstracta (like those listed above) are pretty clear. They are atemporal, non-spatial, Abstract Entities 13 and acausal – i.e., they do not exist in time or space (or space-time), they cannot make anything happen, nothing can affect them, and they are incapable of change. Neither they, their properties, nor events involving them can make anything happen here in the natural world. We don’t see them, feel them, taste them, or see their traces in the world around us. Still, according to a familiar metaphor of some philosophers, they exist “out there,” independent of human language and thought. Being atemporal, non-spatial, and acausal are not all necessary for being abstract in the sense many philosophers have in mind. Thus, many things that seem to be abstract also seem to have a beginning (and ending) in time, among them natural languages like Urdu and dance styles like the Charleston. It may seem tempting to say that such things exist in time but not in space, but where exactly? Moreover, this claim can’t be literally true in a relativistic world (like ours certainly seems to be), where space and time are (framework-dependent) aspects of a single, more basic thing, namely space-time. And not all are suffi cient. For example, an elementary particle (e.g., an electron) that is not in an eigenstate for a defi nite spatial location is typically thought to lack any defi nite position in space. The technicalities don’t matter here; the point is just that although such particles may seem odd, they do have causal powers, and so virtually no one would classify them as abstract. Again, according to many religious traditions, God exists outside of space and time, but he brought everything else into existence, and so many would be reluctant to classify him as an abstract object. All this suggests that the division into concrete and abstract may be too restrictive, or that abstractness may come in degrees. I won’t consider such possibilities here, however, because the puzzles about abstract entities that most worry philosophers concern those entities that are, if they exist, atemporal, non-spatial, and acausal. And we don’t need a sharp bright line between abstracta and concreta to examine these. A philosopher who believes in the existence of a given sort of abstract entity is called a realist about that sort of entity, and a philosopher who disbelieves is called an anti-realist about it. Abstract entities are not a package deal; it is quite consistent, and not uncommon, for a philosopher to be a realist about some kinds of abstract entities (e.g., properties) and an anti-realist about others (e.g., numbers).
Not-quite existence Finally, some champions of abstract entities claim that there are such things, but grant them a lower grade of being than the normal, straightforward sort of existence enjoyed by George Bush and the Eiffel Tower. They often devise esoteric labels for this state; for example, numbers, properties, and the like have been said to have being, to subsist, to exist but not be actual, or partake of one or another of the bewildering varieties of not-quite-full existence contrived by philosophers. Such claims are rarely very clear, but frequently they at least mean that a given sort of entity is real in some sense, but doesn’t exist in the spatiotemporal causal order. Which is pretty much just to say it is abstract. We will not pursue such matters here, however, since many of the same problems arise whether the issue about the status of abstracta is framed in terms of the existence or merely the subsistence or being of such things. Whatever mode of being the number two possesses, we still cannot perceive it, or pick it out in any way, and it seems to Chris Swoyer 14 make no difference to anything here in the natural world. Because many of the most debated issues arise for all the proposed modes of being of abstract objects, I will focus on existence. Why questions about abstracta matter Explicit discussion of abstract entities is a relatively recent philosophical phenomenon. Plato’s Forms (his version of universal properties) have many of the features of abstract objects. They exist outside of space and time, but they seem to have some causal effi cacy. We can learn about them, perhaps even do something like perceive them, though perhaps only in an earlier life (this is Plato’s doctrine of recollection). Soon after Plato, properties and other candidate abstracta – e.g., merely possible individuals (individual things, e.g., persons, that could have existed but don’t) – were reconstrued as ideas in the mind of God. This occurred through the infl uence of Augustine and others, partly under the infl uence of Plotinus and partly under that of Christianity. Human beings were thought to have access to these ideas because of divine illumination, wherein God somehow transferred his ideas into our minds. In later accounts like Descartes’ we had access to such ideas because God placed them in our minds at birth (they are innate). Such views persisted though medieval philoso- phy and well into the modern period. In this period, philosophers like Locke began to view what we thought of above as properties (e.g., redness, justice) as ideas or concepts in individual human minds. It was really only in the nineteenth century, with work on logic and linguistic meaning by fi gures like Bernard Bolzano and Gottlob Frege, that abstract entities began to come into their own. They emerged with a vengeance around the turn of the twentieth century, with work in logic, the theory of meaning, and the philosophy of mathematics, and, more generally, because of a strongly realist reorientation of much of philosophy at this time in the English- and German-speaking worlds. After a few decades, interest in abstract entities subsided, but by the end of the twentieth century, there was perhaps more discussion of a wider array of abstract objects than ever before. Although explicit discussion of abstract entities has a fairly recent history, they are central to debates over venerable philosophical issues, including the nature of mathe- matical truth, the meanings of words and sentences, the features of causation, and the nature of cognitive states like belief and desire. These debates also lie at the center of many perennial disputes over realism and anti-realism, particularly standard fl avors of nominalism. Discussions about the existence of abstract objects may also illuminate the nature of human beings and our place in the world. If there are no abstract objects, nothing that transcends the spatiotemporal causal order, then there may well be no transcendent values or standards (e.g., no eternal moral properties) to ground our prac- tices and evaluations. And if there is also no God, it looks like truth and value must instead be somehow rooted here in the natural order. We are more on our own.
2 Why Believe there are Abstract Objects?
The central questions about abstract objects are: Are there any? If at least some kinds of abstract objects exist, can we discover what they are like? How can we decide such issues? (This question is a problem because it seems to be diffi cult to make contact Abstract Entities 15 with abstract objects in order to learn about their nature.) In this section I will offer an answer to the fi rst question that also suggests an answer to the second. A good way to get a handle on the issues involving abstract entities is to begin by focusing on the point of introducing them in the fi rst place. Philosophers who champion one or another type of abstract object almost always do so because they think those objects are needed to solve certain philosophical problems, and their views about the nature of these abstracta are strongly infl uenced by the problems they think they are needed to solve and the ways in which they (are hoped to) solve them. Hence, our discussion here will be organized around the tasks abstracta have been introduced to perform. These tasks are typically explanatory, to explain various features of philo- sophically interesting phenomena, so to understand such accounts we need to ask about the legitimacy, role, and nature of explanation in metaphysics.
2.1 Philosophical explanations and existence Ontology is the branch of metaphysics that deals with the most general issues about existence. Of course we know a great deal about what sorts of things exist just from daily life: things like trees, cats, cars, other people, the moon. And science tells us more about what sorts of things there are: electrons, molecules of table salt, genes. But ontology attempts to get at the most general categories or sorts of things there are, e.g., physical objects, persons, numbers, properties, and the like. Some philoso- phers doubt that the very enterprise of ontology makes sense (see chapter 9), but we will begin by assuming that it does. For many centuries ontology aspired to be a demonstrative enterprise. On this tra- ditional conception, ontology employs valid arguments to establish conclusions about what the most general and fundamental things in the universe are. It proceeds from obviously secure premises, step by deductively valid step, to obviously secure conclu- sions. The traditional standards for security were very high, requiring unassailable, necessary, self-evident “fi rst principles.” These were supposed to be claims that couldn’t possibly be false and that no reasonable person could doubt. The chief problem with this picture is that when we judge classical arguments in ontology by such standards, most fail, and many fail miserably. There is, among other things, no consensus about which candidates for fi rst principles are even true, much less necessarily so, and, in many cases, demanding valid arguments seems to be asking for too much. By these standards, even the best that the greatest philosophers could devise comes up far short. Nowadays, many philosophers would gladly settle for premises that are uncontro- versially true – or even just fairly plausible. But they still devote a good deal of time distilling arguments for (or against) the existence of one or another sort of abstract object down to a few numbered premises and a conclusion to write on the board; they check for validity, and then (most often) dismiss the arguments. This approach is often invaluable, but it has limitations. For one thing, few philosophical arguments survive long when judged by the pass–fail standards of deductive validity (how likely is it, after all these centuries of inconclusive results, that Jones has just devised an unassailable demonstration that properties exist?). Indeed, it is quite possible that there are no deductively sound arguments beginning from true premises which do Chris Swoyer 16 not mention abstracta and end with conclusions that abstracta exist (“no abstracta in, no abstracta out”). We often miss things of value if we write arguments off simply because they are not deductively valid. But if traditional and contemporary versions of the demonstrative ideal set the bar too high, how should we think about arguments and disagreements in ontology? When we turn to the ways philosophers actually evaluate views about abstract objects, we typically fi nd things turning on the pluses and minuses of one view com- pared to those of its competitors. And a very common feature of the (putative) pluses is that they involve explanation. For example, we are told that the existence of numbers would explain mathematical truth or that the existence or properties (like triangularity) would explain why it is that various objects are triangular and that it would also help explain how we recognize newly encountered triangles as triangles. Moreover, even when the word ‘explain’ is absent, we frequently hear that some phenomenon holds in virtue of, or because of, this or that property, that a property is the ground or foundation or most enlightening account of some phenomenon, or that a property is (in part) the truthmaker, the fundamentum in re (as the medievals would have said) for the phenomenon. For example, it has been urged that the exem- plifi cation of a single, common property grounds the fact that our two items in fi gure 1 (above) are triangular; it makes it true that each is a triangle. The same property also helps to explain how we recognize that they are triangular and why the world ‘triangle’ applies to them. Similar claims have been made on behalf of other abstracta. The role of expressions like ‘explain’ is to give reasons, to answer why-questions, which is a central point of explanation. My suggestion is that we should (re)construe arguments for the existence of abstract entities as inferences to the best overall available ontological explanation (we’ll return to this in sections 3 and 4; see also Swoyer 1982, 1983, 1999a). I will develop this idea in the course of examining the example of numbers, but fi rst let’s see what morals we can draw from the view that arguments for the existence of abstract objects are ampliative (i.e., deductively invalid but capable of offering good, though not conclusive, support for their conclusions). First, we should acknowledge at the outset that there will rarely (probably never) be knock-down arguments for (or against) the existence of any type of abstract entity. On this approach, metaphysics (including ontology) is a fallibilistic, ever-revisable enterprise. By way of example, twentieth-century physics presents us with a very surprising picture of physical reality, and it may well call for innovations in ontology. To note just one case, quantum fi eld theory, that branch of physics that deals with things at a very small scale (quarks, electrons, etc.), strongly suggests that there are (at the fundamental level) no individual, particular things; there may be no fact about how many “particles” of a given kind there are in a particular region of space-time. If so, the traditional view that individuals or substances are a fundamental category of reality may be overthrown. Second, although each specifi c argument for the existence of a certain kind of abstract entity may not be fully compelling, if there are a number of independent arguments that a given sort of entity exists, the claim that they do could receive cumulative confi rmation by helping to explain a variety of phenomena. Abstract Entities 17 Third, if some type of abstract entity is postulated to play particular explanatory roles, this affords a principled way to learn about its nature. We ask what such an entity would have to be like in order to play the roles it is postulated to fi ll. What, to take a question considered below, would the existence or identity conditions of properties have to be for them to serve as the meanings of predicates like ‘round’ or ‘red’? If we are fortunate, we might devise a series of ontological explanations that employ the same entity. This increases information, because different explanations may tell us different things about what that entity is like. It also increases confi rma- tion, because the sequence of explanation may provide cumulative support for the claim that the entity they all invoke actually exists. Explanatory targets and target ranges An explanation requires at least two things. First, something to be explained, an explanation target. Second, something to explain it. In ontology, it is a philosophical theory (though “theory” is often a bit grandiose) like Plato’s theory of forms that does the explaining. We will be concerned with those theories that employ abstract objects in their explanation. Explanation targets for ontology can come from anywhere. From the everyday world around us (e.g., different objects can be the same color, and a single object can change color over time); from mathematics (e.g., it is necessarily the case that three is a prime number); from natural languages (e.g., the word ‘triangle’ is true of many different individual fi gures); from science (e.g., objects attract one another because of their gravitational mass but may repel one another if they are different charges). Explanation targets for ontology can come from almost any area of philosophy (e.g., many moral values seem to be objective, but it’s a bit mysterious how this can be so). I will call a more-or-less unifi ed collection of explanation targets a target domain. In the next section I briefl y discuss several target domains that have led some philosophers to postulate abstract entities. Although I believe that arguments in ontol- ogy are usually best construed as ampliative, much of what follows can be adapted fairly straightforwardly to the view that philosophical arguments should aim to be deductively sound.
3 Examples of Work Abstracta Might Do
When we turn to actual debates about abstract objects, we fi nd few (arguably no) knock-down, iron-clad, settled-once-and-for-all arguments for, or against, the exis- tence of most of the abstract objects that interest philosophers. Instead, the evaluation of the arguments involves the art of making trade-offs, the weighing of philosophical costs and philosophical benefi ts. I will urge that although there are widely shared, quite sensible criteria for this, they fall short of providing rules or a recipe that forces a uniquely correct answer to the question of which, if any, abstract entities exist. Benefi ts rarely come without costs, and we will examine some of the costs of abstracta in section 4. In this section we will consider some of their benefi ts. There are many candidate abstracta and there is space to discuss only one. I will focus on the natural numbers (0, 1, 2, and so on up forever), because this example will be familiar to readers with little background in philosophy. Chris Swoyer 18 3.1 Numbers Target range for philosophy of mathematics There is no unanimity about precisely which mathematical phenomena are legitimate targets for philosophical explanation, but in the case of number theory (basic arith- metic), there is widespread agreement about the following.
1 The sentence ‘7 + 5 = 12’ is true, and its truth is independent of our beliefs and opinions. This is also the case for many other sentences of arithmetic. Similarly, many other sentences of arithmetic, like ‘7 + 5 = 13’ are false, and are so independently of what we happen to think. The truth value (either truth or falsity) is independent of our beliefs and opinions. 2 Statements of arithmetic necessarily have the truth values they do; ‘7 + 5 = 12’ could not have been false under any circumstances and ‘7 + 5 = 13’ could not have been true. 3 Quite apart from questions about language and truth, it is the case that 7 + 5 = 12 but it is not the case that 7 + 5 = 13. And the fi rst is necessarily the case and the second, necessarily, is not. 4 There are infi nitely many natural numbers, and necessarily so (there could not have been fewer). 5 The grammatical structure of the sentence ‘3 is prime’ parallels that of ‘Sam is tall’. In the later case the subject term, ‘Sam’, is standardly thought to denote a real object, the person Sam, and the sentence is true because the thing ‘Sam’ denotes is tall. This suggests that in ‘3 is prime’ the numeral ‘3’ might denote something, and that the sentence is true because the thing it denotes is a prime number. 6 We can employ standard logic in reasoning about arithmetic; the normal, logi- cally valid patterns of inference apply. For example, the step from ‘Sam is tall’ to ‘There is something that is tall’ is a valid inference, both intuitively and in standard systems of logic. So too is the step from ‘3 is prime’ to ‘there is some- thing that is prime’. 7 The claim ‘there is something that is prime’ follows from a true sentence (‘3 is prime’) and seems, quite independently, to be true. But the claim that there is such a thing is just our ordinary, paradigm way of saying that something exists, that it is genuine or really there. Perhaps this is not always the case, but it typically is. So we at least seem to be committed to the view that there is something that is a prime number. 8 It is possible to have reliable justifi ed beliefs and, indeed, knowledge in mathematics. 9 Much of our mathematical knowledge is a priori. This means that we do not need to learn, and almost never justify, our claims in arithmetic by appeal to experience. Once we know what ‘1’ and ‘2’ and ‘+’ and ‘=’ mean, we just see that 1 + 1 = 2.
The list isn’t complete, and some of the items (e.g., 1) may be more central than others (e.g., 2). Still, the more of these targets a philosophical account can explain, Abstract Entities 19 the better. As we will see, however, the features that enable a theory to explain some of these phenomena sometimes make it diffi cult for it to explain others.
Sample explanations in mathematics using abstracta A wide array of philosophical accounts have been developed to explain these targets. I will discuss one of the simplest approaches that employs abstracta. Here is the metaphysical story. The natural numbers are objects or entities, though ones of a very special kind. They are abstract, existing outside of space and time and the causal order. There are infi nitely many of them (what logicians call a denumerable infi nity of them). They do not change. They exist necessarily (they could not have failed to exist), and they necessarily have the properties and stand in the relations that they do (it is necessarily the case that 13 is a Fibonacci number and that 13 > 7). This metaphysical picture allows us to explain item (3) in a very straightforward way. It’s the case that 7 + 5 = 12 because that it just how things are with these mind- independent, objective entities, the numbers – in particular with 7, 5, and 12. And there are infi nitely many natural numbers item (4), because that is just how many of these entities there are (nothing deep here). The purely metaphysical picture may also seem to explain (1) and (2), but to account for matters involving truth, we have to say something about meaning or semantics. Here, as is often the case with accounts of abstract entities, we need to make one or more additional assumptions, auxiliary hypotheses, in order to use those entities to explain the targets we want them to explain. Here we need some semantic auxiliary hypotheses like the following. First, numer- als are singular terms, ones that can occupy subject positions in sentences, and they denote the appropriate numbers (‘0’ denotes 0, ‘1’ denotes 1, and so on out forever). Moreover, numerical terms like these would denote the same things in any possible situation (so they are what philosophers call “rigid designators”). Predicates like ‘prime number’ stand for the property of being a prime number and relational predicates like ‘<’ stand for the relation of being a smaller natural number (I won’t worry here about what these really are). Finally, function expressions like ‘+’ and ‘×’ stand for numerical functions like the addition function (which outputs 5 when you input 2 and 3) and the multiplication function (which outputs 6 when you input 2 and 3). This isn’t the entire story, but it is enough for us to see the basic ideas about how the explanations here work. We then say that a sentence of the form ‘n is P’, where n is a name of a natural number (e.g., the numeral ‘3’) and P is a predicate (e.g., ‘even’), is true just in case n refers to a number that has the property that the predicate P stands for. Similar stories are told for relation and function terms. All of this is a bit loose, but since the work of Alfred Tarski in the 1930s, we know how to make it completely precise. The inter- ested reader can fi nd the details in any good introductory text on symbolic logic, but they aren’t needed to appreciate the basic ideas here. We can now explain why ‘3 is prime’ is true and ‘4 is prime’ is false: ‘3’ stands for an abstract object, the number three; ‘prime’ stands for the property of being prime, and three has that property. By contrast, ‘4’ stands for an abstract object, the number four, that lacks the property. Similar accounts explain why ‘5 < 7’ and Chris Swoyer 20 ‘7 + 5 = 12’ are true and ‘7 < 5’ and ‘7 + 5 = 13’ are false. This explains item (1) on the list. Moreover, since numerical terms necessarily stand for the things that they do, and because the natural numbers necessarily exemplify the properties and stand in the relations that they do, these claims necessarily have the truth values they do (item 2). Simple sentences of arithmetic appear to have a simple subject-predicate structure (item 5; when relation or function terms are involved there is more than one subject term, with ‘5’ and ‘7’ being the two subject terms of ‘5 < 7’). We can now explain this because, given the machinery invoked in our explanations thus far, this is exactly the structure such sentences do have. And we can apply standard logic in a completely straightforward way to explain why normal logical inference rules are valid when we apply them to arithmetical sentences (item 6). For example, existential generalization (the rule that allows us to infer the conclusion “there is something that is F” from the premise “n is F”) works because, if we take a true sentence like ‘3 is prime’, we know that it is true because ‘3’ stands for something (the number three) that is prime. Hence it follows that there is something (three) that is prime. And on this account this sen- tence does indeed make a true existence claim, telling us that there really is something (once again three) that is prime (item 7). Items (8) and (9) differ from the preceding seven insofar as they involve notions like justifi cation and knowledge. These are epistemic notions, ones studied in the philosophical fi eld known as the theory of knowledge or epistemology (from the Greek episteme, ‘knowledge’, and logos, ‘theory’). This is the area of philosophy that deals with knowledge and related concepts like justifi cation. Although this is a different fi eld from ontology, claims about ontology meet up with questions in epistemology when we ask whether, and if so how, we can know about abstract entities. Justifi cation in arithmetic (and in mathematics generally) often proceeds by way of calculations and, at more advanced levels, proofs. These are chains of logically valid patterns of inference. Our previous machinery justifi es the application of logic in arithmetic, and so explains some features of mathematical justifi cation. If we are already justifi ed in believing that ‘3 is odd’, we are then also justifi ed in believing ‘there is something that is odd’. This is so because existential generalization is a mini- valid argument pattern, so if the fi rst sentence is true, the second must be true as well. But our reasoning must begin somewhere. How do we justify those of our arith- metical beliefs that we don’t prove? How do we justify our belief (assuming we take it as basic) that 1 + 0 = 1? Alas, accounts like the one so far that seem well equipped to explain phenomena (1)–(7) founder when we come to (8) and (9). (The classic dis- cussion of this diffi culty is Benacerraf, 1973.) The basic problem is that since numbers are abstract, they lie completely outside the spatiotemporal order. We seem unable to achieve any sort of contact with them. We can’t see numbers, touch them, point to them, measure them. Nor do they cause things we can see or touch or point to or measure. So how do we ever learn anything about numbers? Since all of us know that fi ve is an odd number, we do, somehow, know something about them. On the present account, this knowledge is about an abstract object, namely the number fi ve, though of course a person may not think of it as being an abstract object, perhaps never having heard of such things. But how do we acquire this knowledge? The problem is serious enough that we will defer it in order to treat it in some detail below. Abstract Entities 21 There are competing explanatory accounts of our nine phenomena that employ abstracta other than numbers (especially sets, but also properties, categories, and structures). They have many of the same costs and benefi ts as our simple account using numbers, however, and I will not discuss them here. Finally, we should note that strategies like the one sketched above can be applied in many other parts of mathematics by postulating additional abstract objects, e.g., irrational numbers, complex numbers, and other sorts of mathematical entities (like points, lines, groups, vector spaces).
Lessons the explanations teach us about these abstracta We know a good deal about numbers before we ever study philosophy, so the present philosophical explanations aren’t likely to provide much novel information about their nature (other than telling us that they are abstract). But in the case of less familiar abstracta (like properties or propositions), the explanations might well shed light on the nature of the entity in question. I will say something about what this involves here in the case of numbers to illustrate the sort of thing that is involved in inquires about the nature of any sort of abstract entity. There are at least four things philosophers often want to know about a given sort of entity: its existence conditions, its identity conditions, its modal status, and its epistemic status.
Existence conditions There may not seem to be much philosophical interest in the existence conditions of natural numbers, since we already know which numbers there are (0, 1, 2, 3, . . .; anything you can get by starting with 0 and adding 1 as many times as you like). But with less familiar notions, like that of complex numbers or vector spaces, we typically want to know their existence conditions. Under what conditions is something a complex number? Which (putative) items of that sort exist? The aim is to provide necessary and suffi cient conditions for something being a complex number. To take another example, in set-theory very elaborate conditions are laid down for telling us which sets exist. This model is sometimes carried over from mathematics to philosophy, where phi- losophers ask for the existence conditions for various sorts of non-mathematical abstracta like properties and propositions. It is a matter of debate whether asking for necessary and suffi cient conditions of this sort unreasonably assimilates philosophy to mathematics, but obviously the more their proponents can say about which pro- perties or propositions there are, the better. For example, can there be properties that are not exemplifi ed? Again, assuming that being round and being square are pro- perties, are there also “disjunctive” properties like being round or square, or “conjunc- tive” properties like being round and square?
Identity conditions If x and y are abstract objects, can we provide necessary and suffi cient conditions for them being one and the same object (in the way that 2 and the positive square root of 4 are the same, but 2 and the negative square root of 4 are not)? In the case of numbers, we can typically answer specifi c questions of this sort by calculation or proof, but can we give general identity conditions that apply to all natural numbers in one fell swoop? If x and y have exactly the same numerical proper- Chris Swoyer 22 ties and stand in exactly the same numerical relations, it then turns out that they must be identical, the self-same number. But we might like conditions that throw more light on what it is to be a natural number. By way of example, if x and y are sets, then x and y are identical just in case they contain exactly the same members. Here we get identity conditions that are specifi cally geared to sets (in terms of the notion of set membership), and so are more enlightening about their specifi c nature. Identity conditions are important in mathematics, and as with existence conditions it is possible to worry that requiring identity conditions for a given sort of abstract object as a precondition to granting its existence (or even to discussing whether it exists or not) is an unreasonable demand. After all, philosophers have thus far not been very successful at spelling out precise identity conditions for physical objects or for persons – but we all know perfectly well that physical objects and persons exist. Of course an account of a given sort of abstract object should tell us as much as possible about that object, so we would like to know as much as possible about when x and y are identical, even if this falls short of full necessary and suffi cient conditions for identity.
Modal status Do the abstract entities invoked in our explanations exist necessarily (they simply couldn’t have failed to have existed) or merely contingently (they might not have existed)? Second, we may ask which features of, and relations among, these entities (e.g., being an even number) belong to them necessarily (in any circumstances in which they could exist) and which features only belong to them contingently (they could have existed without having them). Our hope is that if the answers to questions about the nature of a given sort of abstract entity aren’t obvious before developing explanations employing that entity, the explanations themselves will help us answer these questions. In the present case, we hope to see what the modal status of a postulated abstract entity must be in order to explain some of the targets it is supposed to explain. In the explanation sketches of the nine phenomena (listed above on page 19), the answer is that the numbers necessarily exist and that they necessarily have the properties and stand in the relations that they do. We must conclude this in order to explain items (2), (3), and (4).
Epistemic status The most basic epistemological question about an abstract entity we have reason to believe exists is how we can know about it. We can’t reasonably expect a detailed scientifi c answer to such questions at this stage in history, but it would be very useful to be given a general idea. By way of analogy, there is much that we don’t currently understand about visual perception. But we have enough of a general idea how visual perception works to see that it is a normal, natural, causal process involving the refl ection of light off objects to the backs of our retinas, there stimulating nerves and setting off various electro-chemical reactions that, in turn, trigger processes in the visual cortex and other parts of the brain. Admittedly, we don’t understand the conscious aspects of the visual experience itself, but at least they occur in time, surely in space (the brain – besides, you can’t really separate time and space), and involve some sorts of natural, neural causal processes. It would be good to have at least a little detail of this general sort in order to shed light on the way we know about abstract objects. Abstract Entities 23 In the process of answering these questions, we may get an answer to the further epistemic question of whether our knowledge about a given sort of entity (here, natural numbers) is a priori or not. We began by assuming it was, as is traditional, though the account we examined didn’t yield a very satisfying explanation of how this could be so (there are other accounts that do, but they have trouble explaining earlier items on the list). But in the case of at least some other abstract entities, for example properties, there is some debate as to whether our knowledge about them is a priori, i.e., attainable independently of experience (save for enough experience to acquire the concept of them) or a posteriori (based on experience). We might hope to shed light on this question by examining the kinds of jobs that properties are sup- posed to do.
Evaluating explanations in ontology We can rarely explain much with the bald assertions that numbers exist or that prop- erties exist. These claims are typically part of a longer story, a philosophical theory, that tells us something about what the relevant abstract entity is like. The theory also needs to explain how the entity is related to other things, including other abstract entities (theories often invoke more than just one sort of abstract entity, e.g., accounts in semantics often employ both properties and propositions). The account also needs to tell us how its abstracta are related to the phenomena around us that led us to postulate them in the fi rst place. To take a non-mathematical example, a full account of properties should tell us something about which sorts of things have properties (e.g., can properties themselves exemplify properties?), and should help answer questions like whether there exist such properties as colors, shapes, and masses. How are properties related to those things that have them, i.e., what does exemplifi cation amount to? Answers to such questions help us apply the theory of the abstract entity, bridging the gap between the abstract realm and the typically concrete phenomena we want to account for. And an espe- cially important part of an account of abstracta is to tell us at least enough to see that their connection to our cognitive faculties is not hopelessly problematic.
Desiderata There are various desirable features of ontological explanations, features that, other things being equal, make an explanation more compelling.
Do more with less This injunction can take various forms. The fewer unexplained (primitive) entities, the better. If two primitive abstract entities will explain the targets in a domain, don’t use six to do so. The motivation here is general and somewhat vague, but it is important and has a venerable history. The great Medieval philosopher, William of Ockham (c. 1287–1347), counseled philosophers “not to multiply entities beyond necessity.” This precept has become known as Ockham’s Razor, but, as everyone who writes on the matter soon observes, Ockham’s exhortation was to avoid multiplying entities beyond necessity. So the relevant question is always whether a given sort of abstract entity is necessary, which typically means: is it required in order to explain any philosophical targets? The answers to such questions are often Chris Swoyer 24 controversial, so although we can agree that, if a short simple theory works just as well as a long and complicated one, the former is better, in practice, the wield- ing of Ockham’s Razor is often contentious. Breadth and depth of coverage are important The more of the nine arithmetical phenomena (and, indeed, the more additional phenomena) a philosophical theory can explain, the better. Similarly, it counts in favor of a theory of meaning based on properties if it can explain the semantic behavior of different constructions of English (e.g., ‘Sam is tall’ and ‘Sam is taller than Jill’). Explain why rival accounts work as well as they do It is useful if an explanation illuminates why competing accounts work (in those places where they do work) and fail (where they fail). Explain which things need explaining It is also good if an account can illuminate what should, and what should not, be on the list of targets it is used to explain. And if it explains a traditional target away (showing that it doesn’t really exist), it needs to provide arguments for doing so. Don’t solve one problem only to create another just as bad It is important to explain a target (e.g., in semantics) without creating new problems elsewhere (e.g., in epistemology). This list isn’t exhaustive, but it illustrates some of the commonly accepted and central desiderata for explanations in ontology. Unfortunately, these desiderata can be in tension. For example, we can sometimes get by with fewer unexplained entities by sacrifi cing breadth of coverage. Hence, these goals do not add up to rules or recipes that always tell us which of several competing philosophical explanations is best, and this remains the case even if we add further plausible desiderata. But it is important that they often can tell us that certain explanations are not very good. Constraints There are also constraints on ontologically satisfactory explanations. Some are nearly universal (e.g., logical consistency, though even that has been challenged lately). Others vary with time or schools of thought, and some refl ect quite idiosyncratic philosophical scruples or ideals. Constraints and desiderata fade into one another, but the importance of the former should not be underestimated. For example, in various periods there have been religious constraints on metaphysi- cal explanations. In medieval disputations about properties, issues involving faith, reason, and the nature of God were never far from view. Indeed, these matters often provided explanatory targets for metaphysicians. Philosophical orientations also provide constraints. For example, many philosophers have argued that knowledge must be grounded in experience. We cannot simply reason out what the world is like from the armchair; we have to go and check. Today, naturalistic world-views are popular, and these are often thought to allow only physical entities, or at most only entities that exist in space-time. In concrete historical settings, constraints can seem very real, sometimes inevitable, even if at a latter time they seem arbitrary, even quaint. This needn’t make metaphys- ics “subjective” in any debilitating sense (so that whatever a particular culture happens to think about it is “true for them”). But it is a useful reminder that metaphysics, like Abstract Entities 25 any other intellectual enterprise, is a human endeavor that takes place in, and is highly colored by, a time, culture, and tradition.
Diffi culties with competitors The best available ontological explanation must meet some minimal threshold of goodness to justify belief in its conclusion. Moreover, the notion of the best available explanation is comparative; a theory doesn’t get many points for explaining something if a rival theory explains it much better. Hence, arguments for the existence of a particular sort of abstract entity often need to be bolstered by criticisms of opposing theses. For example, the view that properties are needed to play the role of semantic values of predicates is stronger when accompanied by arguments that other sorts of entities, e.g., sets (of the things to which the predicate applies), cannot play the role nearly as well. Finally, being the best explanation doesn’t mean being perfect. Virtually all philosophical accounts have open problems, and trying to solve them is part of the day-to-day work of philosophers.
Quandaries and doubts I have spoken as though inference to the best ontological explanation were relatively unproblematic, but there are various places where objec- tions to it, and especially to its use in philosophy, can be raised. I have discussed these matters elsewhere (e.g., Swoyer 1983; 1999a; 1999b), however, and will not pursue the matter here.
4 Pluses and Minuses
Metaphysics, like life, often presents us with diverse, not fully compatible, goals that require us to make trade-offs, weigh costs against benefi ts, make hard decisions. In this section we consider the chief benefi ts, and costs, of abstract entities.
4.1 Benefi ts The primary philosophical attraction of abstract entities is that they seem to offer so much explanatory power. For example, when we encounter words or phrases that look like denoting singular terms (e.g., ‘3’, ‘courage’) we can explain this very neatly by arguing that they are singular terms and that they denote an abstract entity (a number, a property). The realist can often avoid denying the existence of relatively obvious phenomena (like the existence of mathematical truth, which some anti-realists about numbers and sets deny), needn’t urge that we have been badly in error about entire realms of discourse (like mathematics), and can avoid resorting to tortured paraphrases to evade ontological commitment. Indeed, the more luxuriant lines of abstracta (e.g., Russell 1903; Zalta 1988; Bealer 1982) contain so much metaphysical machinery that it is almost a foregone conclusion that they can explain any phenom- enon that comes their way. All this sounds a little to good to be true. Is it?
4.2 Costs Ockamist impulses and ontological economy Few philosophers like ontological bloat. Other things being equal, a good explanation of a philosophical target that doesn’t rely on abstracta is preferable to a good expla- Chris Swoyer 26 nation that does. But other things are rarely equal. Abstract objects often add enough explanatory power that theories invoking them can give broader and smoother expla- nations of a target than theories that do not. For example, it is very diffi cult (though a number of philosophers believe not impossible) to give an account of mathematical truth that does not employ abstracta of any sort. While ontological economy is impor- tant, other things are rarely equal, so it is rarely decisive.
Anti-realism: there are alternatives There is always anti-realism, so perhaps we shouldn’t feel driven to abstracta as the only game in town. There are many forms that opposition to realism takes nowadays, as new positions (e.g., fi ctionalism, projectivism, error theories) spill over from the philosophy of mathematics, the philosophy of science, and meta-ethics into philoso- phy generally. Furthermore, with the demise of behaviorism, philosophy’s linguistic turn is beginning to show its age, and the rise of cognitive science, and various fl avors of conceptualism, are once again on the menu (e.g., Swoyer 2005). Still, none of these alternatives provides a strong reason for avoiding the need for a given sort of abstract object to explain a legitimate philosophical target unless the anti-realist explanation (or dismissal) of it is spelled out in a reasonably detailed and compelling way. So again, we must consider each approach case by case.
Epistemic access Epistemology is the Achilles’ heel of realism about abstracta. We are biological organisms thoroughly ensconced in the natural, spatiotemporal causal order. Abstract entities, by contrast, are atemporal, non-spatial, and causally inert, so they cannot affect our senses, our brains, or our instruments for measuring and detecting. A few philosophers have postulated a cognitive faculty of intuition that provides some sort of non-causal access to numbers or other abstracta. The nature of this access has never been explained, however, and many of us fi nd nothing like it in our own perception and thought. Scientists have no inkling where it is located in the brain, and it has yet to turn up in any empirical studies. Empirical investigation of thought that (might) seem to be about abstracta is becoming more common (e.g., Boroditsky and Ramscar 2002), and it may eventually illuminate the issues here. At present, however, it doesn’t get at the most basic problems that have worried philosophers about our cognitive access to abstracta. Perhaps knowledge about abstracta doesn’t require contact with them. The only remotely plausible story about this would seem to be that such knowledge is innate. This may well be true of our rudimentary knowledge of arithmetic, but it doesn’t scale up well to knowledge about tensor algebra or the semantic values of words for describ- ing the nuances of medieval chivalry. The epistemic problems here do not stem from any (almost certainly hopeless) causal theory of knowledge, but simply from the fact that our acquisition and justi- fi cation of beliefs about things lying outside the spatiotemporal causal order is more than a little mysterious. Indeed, even if abstracta did exist, it is diffi cult to see how they could make any difference to our cognitive processes. Things would seem just the same whether they existed or not, or if they existed up until tomorrow, then suddenly vanished. Abstract Entities 27 Reference and non-uniqueness Nowadays a major reason for postulating abstracta is to use them as semantic values in semantic accounts of natural languages. Unfortunately, the epistemic problems abstracta generate make it diffi cult to use them for this purpose. We can’t make epistemic contact with abstracta, so it is diffi cult to see how we could get our words to latch onto them. We can’t single numbers out, by pointing or in any other obvious way, and say ‘that is 0’, ‘that is 37’, and so on. We might try to pick 0 out by saying that ‘0 is the fi rst of the natural numbers’, but this doesn’t really help unless we have pinned down the reference or extension of ‘natural number’ and (less obviously) that of ‘fi rst’ (as it applies to the sequence of natural numbers). So we are back with the original problem. We can’t make identifying reference in language because we can’t make identifying reference in thought. In some cases, particularly in mathematics, we can specify the structure of a given realm of abstract entities. For example, we can pin down the “structure” of the natural numbers with some sophisticated logic (with what is known as a second-order version of Peano’s Postulates). The structure is roughly this: there is a fi rst object in the structure (0); there is a second object in the structure (1) with nothing in between it and the fi rst object; there is a third object (2) with nothing in between it and the second object; and so on forever. That is, the structure is that of an infi nite, discrete series with a beginning but no end. But if there is one group of things with this structure, then, logicians have demonstrated, there are many, and there is little reason to suppose that any one of them gives the unique metaphysical truth about “What Numbers Really Are” (cf. Benacerraf 1965). Because we lack epistemic contact with numbers, we can only describe the structure of the realm of numbers, and such descriptions underdetermine the denotations of our numerical vocabulary. So, ironi- cally, the apparent success of our earlier explanations for the semantic features of numbers seems undermined by the problems with epistemic phenomena.
5 Conclusion
So . . . are there abstract entities? And if so, which ones? The answers depend on the answers to three prior questions. Is inference to the best available overall ontological explanation ever legitimate? If so, when? And when it is, how do we adjudicate among competing explanations? My answers are more tentative than I would like, but this is a conclusion, so I will end by drawing some. Is the game optional? If someone won’t play the metaphysical game, there are no knock-down, non-ques- tion-begging arguments to show she is wrong. We can cite reasons for, and against, the possibility of inference to the best ontological explanation, but none of them comes close to being conclusive. Indeed, if I am right, differences of beliefs in ontol- ogy very often stem from differences of beliefs about the legitimacy and nature of inference to the best explanation in ontology. Evaluating competing explanations The gist of the discussion thus far is that evaluation of rival explanations in ontology is a global affair that requires sound philosophical judgment rather than a reliance Chris Swoyer 28 on hard and fast rules (the problem is that there are no such rules, though there are rough but generally accepted guidelines, so that not just anything goes). The process is global or holistic, in the sense that it depends on the weighing of many different considerations at the same time. And although the decisions that must be made in evaluating competing programs are usually made in light of shared philosophical values, there doesn’t seem to be any uniquely correct way to trade such values off against each other. For example, other things being equal, more explanatory power, breadth of coverage, and simplicity are better than less. But then, when are things ever equal? And when they are not, is it better to have a richly detailed explanation of a narrower range of phenomena or a less detailed explanation of a wider range?
Disagreements about simplicity Arguments over simplicity play a prominent role in debates in ontology, sometimes crowding out consideration of other important explanatory virtues. The verdict of simplicity is rarely unequivocal, however, and judgments about it differ from one philosopher to another. Still, some philosophical disputes actually come down in print to questions about whether two basic, undefi ned, primitive objects and one basic, undefi ned, primitive relation are simpler than one primitive object and two primitive relations. Such considerations are surely much too fragile to support conclusions about the “ultimate nature of reality,” as if “What There Really Is” could come down to whether an account employs two primitive notions, rather than three.
The fundamental ontological trade-off There is an even more fundamental trade-off that we face at every turn in philosophy, from ethics to philosophy of science to philosophy of mathematics to metaphysics. I will call it the fundamental ontological trade-off. This is the trade-off between explanatory power, on the one hand, and epistemic credibility, on the other; between a rich, lavish ontology that promises a great deal of explanatory punch, and a more modest ontology that promises more epistemological security and believability. How a philosopher strikes a balance in this trade-off goes a long way to determining whether or not she will believe there are abstract entities. The more machinery (especially abstract machinery) we postulate, the more we might hope to explain – but the harder it is to believe in the existence of all that machinery. Russell makes this sort of point in his famous theft-over-honest-toil passage: “The method of ‘postulating’ what we want has many advantages; they are the same as the advantages of theft over honest toil. Let us leave them to others and proceed with our honest toil” (Russell 1919: 71). But without at least a little postula- tion, it is very diffi cult to even get started.
The upshot Once ontological explanation is allowed and (rough-and-ready) ground rules are set, there can be winners and losers and perhaps a spectrum of views in between, but it is important that not everyone who plays the metaphysical game gets to win. For example, although Goodman and Quine’s (1947) celebrated attempt to provide an account of mathematics that avoided all abstracta remains impressive, it simply cannot account for enough features of mathematics to be judged a success – even by Quine. Abstract Entities 29 But once we eliminate the more unpromising explanations, we may well be left with more than one contender. In short, if ontological explanations are legitimate, it is unlikely that there will be uniquely correct explanations, and so unlikely that we will arrive at a single picture about which abstracta (if any) there are. Perhaps we can make slow progress to this goal, with a series of explanations zeroing in more and more on the existence and nature of various abstracta. But such a series may instead lead to a fragmentation of entities, with a corresponding fragmentation of our views about them.
If epistemology isn’t a problem, then abstracta win If inferences to the best ontological explanation are legitimate, and if the epistemic problems about cognitive access to abstract entities can be overcome, then the case for at least some abstract entities is very strong. This is so because we can explain much more with them than without them. But the epistemological problems are severe.
A parting thought Still, would it really be so bad if the best we could do was to rule out some accounts in ontology and learn to live with more than one survivor? Perhaps developing a tolerance for more than one (which need not mean every) ontological framework is the best we can do. If we can do this without falling into some dreadful sort of rela- tivism, maybe that’s good enough.
Note
1 Many discussions of abstract objects are rather technical, but in the interests of accessibility I will steer clear of such complexities and avoid logical notation (the interested reader can fi nd many of the more technical matters discussed in some of the works cited here). Because the existence of various sorts of abstract objects, and indeed abstract objects in general, is a matter of contention, prudence suggests constant qualifi cations like “putative” examples of abstract entities and talk that “seems” to be about them. But this becomes tiresome and I will mostly leave such hedges tacit. I am grateful to David Armstrong, Hugh Benson, Monte Cook, Brian Ellis, Ray Elugardo, Jim Hawthorne, Herbert Hochberg, Chris Menzel, Adam Morton, Sara Sawyer, Ted Sider, Shari Villani, and Ed Zalta for helpful discussions on the topics discussed here.
References
Bealer, George. 1982. Quality and Concept (Oxford: Clarendon Press.) Benacerraf, Paul. 1965. “What Numbers Could Not Be,” Philosophical Review 74: 47–73. ——. 1973. “Mathematical Truth,” Journal of Philosophy 70: 661–79. Boroditsky, Lera, and Ramscar, Michael. 2002. “The Roles of Body and Mind in Abstract Thought,” Psychological Science 13: 185–9. Goodman, Nelson, and Quine, W. V. O. 1947. “Steps Toward a Constructive Nominalism,” Journal of Symbolic Logic 12: 97–122. Chris Swoyer 30 Russell, Bertrand. 1903. Principles of Mathematics (New York: W. W. Norton & Co). ——. 1919. Introduction to Mathematical Philosophy. London: Routledge; repr. 1993. Swoyer, Chris. 1982. “The Nature of Natural Laws,” Australasian Journal of Philosophy 60: 203–23. Repr. in Theory, Evidence and Explanation, Darthmouth Publishing Co., 1995. ——. 1983. “Realism and Explanation,” Philosophical Inquiry 5: 14–28. ——. 1999a. “How Metaphysics Might be Possible: Explanation and Inference in Ontology,” Midwest Studies in Philosophy: New Directions in Philosophy 23: 100–31. ——. 1999b. “Properties,” in Edward Zalta, ed., Stanford Encyclopedia of Philosophy; available at
Abstract Entities 31 CHAPTER 1.2
There Are No Abstract Objects
Cian Dorr
1 The Thesis
Suppose you start out inclined toward the hard-headed view that the world of mate- rial objects is the whole of reality. You elaborate: ‘Everything there is is a material object: the sort of thing you could bump into; the sort of thing for which it would be sensible to ask how much it weighs, what shape it is, how fast it is moving, and how far it is from other material objects. There is nothing else.’ You develop some practice defending your thesis from the expected objections, from believers in ghosts, God, immaterial souls, Absolute Space, and so on. None of this practice will do you much good the fi rst time you are confronted with the following objection:
What about numbers and properties? These are obviously not material objects. It would be crazy to think that you might bump into the number two, or the property of having many legs. One would have to be confused to wonder how much these items weigh, or how far away they are. But obviously there are numbers and properties. Surely even you don’t deny that there are four prime numbers between one and ten, or that spiders and insects share many important anatomical properties.1 These well-known truths evidently imply that there are numbers, and that there are properties. So your thesis is false. Not everything is a material object.
This disconcertingly simple objection is probably quite unlike anything you expected to have to deal with when you fi rst announced your thesis. It is confusing precisely because it is so very simple: if the argument did lead you to give up your initial materialist beliefs, the fact that you ever held those beliefs in the fi rst place should seem profoundly puzzling. How on earth could you have failed to notice the incon- sistency between your belief that spiders and insects share many important anatomical properties and your belief that everything is a material object? Seeing this, you will quite naturally wonder whether your disagreement with the objector might not be merely verbal. You may want to begin your reply by making distinctions: ‘Of course there is in a sense such a thing as the number two; but in another important sense it is still true that material objects are all there are.’ While I have no particular interest in defending the view that the world of material objects is the whole of reality, I think that this reply is right on target. Sentences like
(1) There are four prime numbers between one and ten. (2) Spiders and insects share many important anatomical properties. (3) There are numbers. (4) There are properties. (5) Everything is a material object. admit of (at least) two systematically different kinds of uses, which I will label “fun- damental” and “superfi cial”. Ordinary uses of sentences like (1) and (2) are superfi cial, and entirely appropriate. When we use these sentences superfi cially, we assert boring, well-known truths, just as we would have if we had said that spiders and insects both have exoskeletons, or that there are no square prime numbers. On the other hand, anyone who seriously uttered (5) would very likely be using this sentence in the fun- damental way. They would be making a claim about “the ultimate furniture of reality”: the claim that in the fi nal analysis, there are only material objects. This claim is per- fectly consistent with the truths that would be expressed by ordinary, superfi cial uses of sentences like (1) and (2). The appearance of confl ict arises from the fact that each of the above sentences can be used in both ways, although the two uses are not always equally natural. (3) and (4) can be used superfi cially, in which case they would express truths even easier to know than the ones expressed by superfi cial uses of (1) and (2). Even (5) could in principle have superfi cial uses, which would express some- thing obviously false. Conversely, (3) and (4), and even (1) and (2), can in principle be used in the fundamental way, in which case they would express claims that are far from obvious, and inconsistent with the claim expressed by fundamental uses of (5). Thus, the arguments from (1) to (3), from (2) to (4), and from (3) or (4) to the negation of (5) are valid in the sense that the conclusions follow from the premises provided they are used in the same way. I hope this all seems too obvious to need arguing for as opposed to pointing out. How strange it would be to think that ordinary people, including those who never give a thought to questions of metaphysics, hold a wide range of beliefs that are incompatible with the thesis that, fundamentally speaking, everything is a material object, and express these beliefs whenever they use sentences like (1) and (2)! But if an argument is wanted, I can offer the following. In ordinary life, we treat arguments such as these as trivially valid:
(6) (a) There is a planet that is distinct from some planet. (b) So, the number of planets is greater than one. (7) (a) The Earth is round. (b) So, the Earth has the property of being round. There Are No Abstract Objects 33 Indeed, it is easy to hear the conclusions of these arguments as nothing more than stylistic variants, or “pleonastic equivalents” (Schiffer 2003), of the premises. Surely this can’t be an outright mistake. But, taken in the fundamental sense, (6b) and (7b), like (1) and (2), express substantive metaphysical claims – claims that would be false if there were, fundamentally speaking, only material objects. These controversial claims certainly do not follow analytically – just as a matter of meaning – from (6a) or (7a). Whether we take (6a) and (7a) themselves as fundamental or as superfi cial, there is manifestly nothing in their meaning that could stop them from being true if the world of material objects were the whole of reality.2 I conclude that (6b) and (7b) are standardly used non-fundamentally, to express something consistent with the claim that there are, in the fundamental sense, no numbers or properties. But if this is granted for (6b) and (7b), it must also be granted for (1) and (2). Clearly we don’t take any more of a metaphysical risk in uttering the latter sentences than we do in uttering the former: in whatever sense our utterances of (1) and (2) commit us to the existence of numbers and properties, our utterances of (6b) and (7b) do the same. Have I said enough by now to enable me to state my thesis without being misun- derstood? Let’s give it a try. There are no numbers. There are no properties. When I utter these sentences, I mean to be using them in the fundamental way. I mean, if you like, that numbers and properties are not part of the ultimate furniture of reality. I mean that there are, in the fi nal analysis, no such things. Of course I hold similar views about many of the other putative entities that generally get classifi ed as “abstract”. But I don’t want to be drawn into a pointless debate about how to defi ne that technical term, so I won’t attempt to state any more general thesis from which these claims about numbers and properties could be derived.3 I will, however, add two more claims at the same level of generality. Just as there are no numbers or properties, there are no relations (like being heavier than or betweenness), or sets (like the set of people who have read this paper, or the null set). I will provisionally use ‘nominalism’ for the conjunction of these four claims, but will try to say nothing involv- ing this term that would not also be true of various stronger theses of a similar kind. ‘So are sentences like (1), (2), (3), and (4) true or false, according to you?’ Well, just as these sentences admit of divergent fundamental and superfi cial uses, so too do sentences like (1*) ‘There are four prime numbers between one and ten’ is true. (2*) ‘Spiders and insects share many important anatomical properties’ is true. (3*) ‘There are numbers’ is true. (4*) ‘There are properties’ is true. and their negations. Perhaps fundamental uses of (1*) and (2*) and their negations are a bit less unnatural than fundamental uses of (1) and (2). If you heard someone utter (1*) or (2*), you would naturally wonder why they didn’t just assert (1) or (2); one reasonable hypothesis is that they did so as a signal that they were speaking fundamentally. But one wouldn’t want to rely on hearers to come up with this hypoth- esis on their own.4 ‘Yes, but which of the above sentences are literally true?’ Well, I’m certainly not play-acting, or speaking metaphorically, or exaggerating, or being sarcastic, when in Cian Dorr 34 the course of doing philosophy I say things like ‘there are no properties’, or when in the course of doing other things I say things like ‘spiders and insects share many impor- tant anatomical properties’. So if “literally” is used in the ordinary way to rule out these kinds of verbal maneuvers, I’ll happily say in the fi rst kind of context that there are literally no properties, and in the second that spiders and insects literally share many important properties. However, “literal” has come among linguists and philoso- phers of language to have an extended use, on which it is supposed to stand for some category of deep explanatory importance: the domain of semantics rather than prag- matics. A way of using language might turn out to be non-literal in this sense if it turned out, at some deep explanatory level, to involve the same kinds of capacities or mechanisms that are involved in metaphor, exaggeration, sarcasm, etc. It has occasionally been suggested that what I have been calling the “superfi cial” uses of sentences like (1)–(5) are non-literal in this sense.5 But I have no interest in arguing for this surprising empirical claim. Indeed, I wouldn’t mind if the fundamental use of sentences like (1)–(5) turned out to be non-literal in this sense. Why should it matter, provided I can get my meaning across? Since I think that the claims expressed by the fundamental uses of sentences like ‘there are numbers’ and ‘there are properties’ are of great and enduring philosophical interest, I am naturally disposed to assume that philosophers who utter these sentences mean to be asserting these interesting claims, rather than the trivialities expressed by these sentences on a superfi cial interpretation. But in some cases, there is other evi- dence that makes the assumption quite problematic. I am thinking especially of those philosophers who produce these sentences as the conclusions of arguments such as the following:
(6) (a) There is a planet that is distinct from some planet. (b) So, the number of planets is greater than one. (c) So, there are numbers. (7) (a) The Earth is round. (b) So, the Earth has the property of being round. (c) So, there are properties.6
As I argued above, these arguments are only valid on a superfi cial reading. The fact that someone puts forth such an argument as valid is thus weighty evidence in favor of an interpretation on which that person is speaking superfi cially. On the other hand, why would any philosopher court misunderstanding in this way? In some cases I suppose the right explanation is that, by failing to be clear about the distinction between the two uses, they have ended up believing that there are numbers or properties in the fundamental sense on the basis of fallacious arguments. But in the more interesting cases, the explanation is, rather, that the philosophers in question see no alternative to the superfi cial way of using sentences like ‘there are numbers’. They simply have no idea what the allegedly distinct “fundamental” uses of these sentences are supposed to be. (Or at least, they have convinced themselves that they have no such idea.) Here’s how I imagine them responding to my thesis:7
Look, I’d like to interpret you charitably as not expressing that claim – the one I would express using the words ‘there are no numbers’ – which we agree is a trivial analytic There Are No Abstract Objects 35 falsehood. But I just don’t see what else you could have in mind. The only alternative interpretations of your utterance I can think of are ones on which you are expressing some trivial analytic truth (e.g., that numbers are not concrete), or some empirical claim to which your arguments are plainly irrelevant (e.g., that it would be in our interest to adopt a practice in which ‘there are no numbers’ was used to express a truth – cf. Carnap 1950). I am in a position like the one you would be in if you were faced with someone who maintained that the objects we normally refer to as “chairs”, although they do exist, are not really, strictly speaking, in the fundamental sense, chairs. You can’t see what could make it appropriate to dignify any variant on our ordinary use of the word ‘chair’ as especially “strict” or “philosophical” or “fundamental”. I am in just the same position with regard to the ordinary use of ‘there are’ – the one on which ‘there are numbers’ is used to express a trivial truth.
I would be delighted to be able to argue people out of this position of principled incomprehension, by fi nding a distinguishing characteristic of the “fundamental” use of expressions like ‘there are’ that even they would have to recognize. My hope is that this could be done by taking my claim that arguments like (6a)–(6b) and (7a)–(7b) are not valid when understood in the fundamental sense, and the related claim (note 2) that claims of existence are never analytically true in the fundamental sense, as partly defi nitive of the relevant use of ‘in the fundamental sense’. But I won’t attempt here to develop this into a workable characterization.8 For now, let me simply point out that the two anti-nominalist arguments I will spend most of this chapter discuss- ing (sections 3 and 4) purport to show that abstract objects do essential explanatory work of some sort. If these arguments succeed, any way of using language on which ‘there are no abstract objects’ was used to express a truth (while ‘abstract object’ continued to mean what it actually means) would be severely explanatorily impover- ished. Even those who can make no sense of the claim that such a way of using language is “more fundamental” than our ordinary way of using language might want to resist this claim. They might think that the fact that ‘there are abstract objects’ is true in ordinary English is just a fact about how we happen to talk, rather than something forced on us by our interest in speaking a language that is not explanatorily impoverished.9 Those who hold this view have a common interest with me in fi nding responses to the anti-nominalist arguments I will be discussing.
2 Paraphrase
The superfi cial way of talking about numbers, properties, relations, and sets is very useful. But it would be wrong to think of it as giving us access to a domain of inde- pendent fact from which we would otherwise be completely cut off. Rather, sentences get to be true or false taken superfi cially in virtue of what there is in the fundamental sense, and what it is like. Thus, each English sentence must have a “paraphrase”: a sentence that, when taken in the fundamental sense, says how things would have to be for the original sentence to be true in the superfi cial sense.10 For some sentences, appropriate paraphrases are ready to hand. For example, we could take (6a) (‘There is a planet that is distinct from some planet’) as the paraphrase of (6b) (‘The number of planets is greater than one’).11 But there is no obvious way to generalize this Cian Dorr 36 assignment beyond these easy cases. If it turned out that no system of paraphrases capturing the apparent logical relations among superfi cial uses of English sentences was possible, that would undermine my case for the existence of non-equivalent fundamental and superfi cial interpretations of abstract-object sentences.12 There would then be no choice but to reject at least one of the premises to which I appealed in arguing for that claim – for example, that we are not making a mistake when we treat arguments like (6a)–(6b) and (7a)–(7b) as valid.13 If we wanted to continue to be nominalists, we would have to do so in the knowledge that we were disagreeing with something almost everyone has believed. And when we ourselves continued to talk in the usual way about numbers, properties, relations, and sets (as we inevitably would), we would fi nd it hard to explain what we were doing if not expressing beliefs fl atly inconsistent with our offi cial doctrine. Fortunately, the challenge to provide the required system of paraphrases can be met rather easily. We can simply take the paraphrase of any sentence S to be ‘If there were abstract objects, it would be the case that S.’14 Or, if one would prefer the para- phrases to be more explicit and less context-dependent, one could aim to fi ll in the following schematic analysis:
If it were the case that [axioms of number-theory] and [axioms of set-theory] and [axioms of property and relation theory] and [axioms for other sorts of abstract objects], and the concrete world were just as it actually is, then it would be the case that S.15
One objection to this sort of proposal stems from the widely accepted thesis that counterfactual conditionals with metaphysically impossible antecedents are all vacu- ously true (Stalnaker 1968, Lewis 1973). Since nominalism seems like the sort of thing that should be metaphysically necessary if true, this thesis would be bad news for my proposal. Fortunately, the thesis is false (see Nolan 1997). Here are some manifestly false counterfactuals whose antecedents seem to be metaphysically impossible:
(8a) If I were a dolphin, I would have arms and legs. (8b) If it were necessary that there be donkeys, it would be impossible that there be cows. (8c) If there were unicorns, none of them would have horns.16
In fact we seem to be just as good at fi nding sensible things to say about what would be the case if some impossibility were true as we are at understanding what is the case according to impossible fi ctions.17
3 Abstract Objects in Scientifi c Explanations
If this reply to the “paraphrase challenge” succeeds, the picture from section 1 stands. There are plenty of short and valid arguments from obvious truths to the conclusion that there are numbers and properties in a superfi cial sense; but such arguments are irrelevant to the question whether there are any such things in the fundamental sense. It seems clear to me that in that case, Ockham’s Razor applies: the burden of proof There Are No Abstract Objects 37 lies with those who maintain that there are, in the fundamental sense, numbers, properties, relations, or sets. Unless they can show us that their view brings some important explanatory benefi t that cannot be had more cheaply, we should be reason- ably confi dent that there are no such things. We should regard it as far more likely that all the things that there really are are material objects, or spirits, or portions of space-time – or at any rate, that they are all causally or spatiotemporally or mentally or physically interrelated as no number, property, relation, or set could ever be.18 The disjunctive character of this claim makes it hard to say anything further at this level of generality to justify it. If further justifi cation is demanded, the nominal- ist’s best strategy may involve arguing for some more specifi c, positive thesis about what reality is like, fundamentally speaking – something strong enough to have explanatory power in its own right, and from which nominalism could be derived as a corollary. I won’t try to do this here. Instead, I will spend the remainder of the chapter considering two of the ways in which anti-nominalists have attempted to put abstract objects to work in explanations. I agree with Chris Swoyer (see chapter 1.1) that if it could be shown that the thesis that abstract objects exist (fundamentally speaking) is required by some good explanation of something that couldn’t otherwise be explained, or explained so well, that would provide at least some reason to believe it. I will argue, however, that the explanatory benefi ts abstract objects may seem to bring are in fact illusory. In this section, I will consider whether the reality of abstract entities might be required for good scientifi c explanations – explanations that answer ‘why’ questions in the way in which Newton’s laws answer the question why the planets move in elliptical orbits. In section 4, I will turn to the question whether the reality of abstract entities might be required for good philosophical, or constitutive, explanations – explanations that answer ‘why’ questions in the way in which the claim that a certain object is a four-sided fi gure with equal sides and angles answers the question why it is a square.19 The idea that belief in certain abstract entities is justifi ed by their role in good scientifi c explanations is called the “indispensability argument”, and generally attrib- uted to Quine (1948) and Putnam (1971). At fi rst sight, the case looks strong, especially for mathematical entities. Why would modern science be so full of theories that logi- cally entail ‘there are numbers’ if everything these theories purport to explain could be explained equally well without positing mathematical entities? But on second thought, it seems quite easy to weaken any ordinary scientifi c theory so as to make it logically consistent with ‘there are no numbers’ without affecting the theory’s pre- dictions about the concrete world. We already saw one way of doing this in section 2: when T is an ordinary scientifi c theory incorporating certain mathematical axioms,
T* If it were the case that [mathematical axioms] and the concrete world were just as it actually is, it would be the case that T has exactly the same consequences for the concrete world as T itself. So the proponent of the indispensability argument owes us an argument that T* provides a worse explanation for these facts about the concrete world than T does. Why would one think this? Cian Dorr 38 Some anti-nominalists appeal at this point to the standards actually accepted by practicing scientists. They point out that one would get short shrift if one were to submit something of the form of T* to a journal like the Physical Review (Burgess and Rosen 1997: 210). Philosophers, they suggest, should defer to the wider scientifi c com- munity in forming their opinions about what makes for a good explanation. But if I am right that most uses of abstract-object talk, even in the sciences, are superfi cial uses, this sort of direct appeal to authority is useless in an argument for the existence of mathematical entities in the fundamental sense.20 True, scientists never pause to con- sider theories looking like T*, but that’s because the theories they are interested in are already of this form, under the surface. Thus, any guidance scientists can give us in our attempt to assess the explanatory goodness of theories like T* must be indirect: it must depend on some analogy between this case and other cases where scientifi c practice clearly does take a stand on what is required for a good explanation. It is not hard to see how such an argument by analogy might go. Putnam (1971) is naturally interpreted as giving an argument of this kind; let me present my own version. If we are “scientifi c realists” who accept science as a way of fi nding out about the world, including the unobservable portions of the world, we will presumably think that we have good reason to believe that there are subatomic particles.21 We will think this even though we recognize that, for any theory T that talks about unobservable entities of some kind, it is possible to fi nd a weaker theory that shares all T’s consequences about the observable world without entailing anything at all about those unobservable entities. For example:
T † All the facts about the positions and motions of atoms are consistent with the hypothesis that T.22
(In other words: as far as the positions and motions of atoms are concerned, it is just as if T were true.) How can we be justifi ed in believing that there are subatomic particles, rather than cautiously limiting ourselves to beliefs of the form of T †? The standard answer makes use of the notion of explanation. Although T † has exactly the same observable consequences as T, it does not constitute a good explanation of these consequences. For T † is the sort of theory which, if true, would itself “cry out” for further explanation. It is extremely unlikely, a priori, that atoms should just happen to move around in exactly the ways predicted by T without this fact having any deeper explanation. T itself is the most obvious possible explanation. There may of course be other explanations inconsistent with T, including some we have not yet thought of. But it would be very surprising if someone were to discover a theory that would, if true, provide a good explanation of T † without entailing the existence of subatomic particles. Now, T* and T † are similar in some salient respects. There is an obvious sense in which both are “parasitic” theories. Once the original theory T has been written down, it requires no additional effort to formulate the weaker theories T* and T †; one simply prefaces T with the appropriate complex modal operator, ‘If [mathematical axioms] and the concrete world were just as it actually is, it would be the case that . . .’ or ‘The facts about the positions and motions of atoms are consistent with the hypothesis that. . . ’. On the basis of this similarity, one might conclude that T* and T † must also There Are No Abstract Objects 39 be similar in respect of being bad as explanations – i.e. in being unlikely to be true without having any deeper explanation.23 Once it has been made explicit, this argument by analogy may seem too tenuous to be worth worrying about. T* and T † are dissimilar in all sorts of ways. Why should we rest so much on the ways in which they are similar? But on the other hand, what else are we to go on, besides such analogies, in assessing the merits of T*? It is hard to think of any other way of deciding the status of T* that would not beg the ques- tion by appealing to some premise that only nominalists or anti-nominalists fi nd plausible. For example, many nominalists have seen epistemological signifi cance in the fact that abstract objects, if they existed, would be causally inert (Benacerraf 1973). The number two is not the sort of thing that could, say, move a pointer on some properly designed number-detector. Appealing to this fact, a nominalist might attempt to argue that T* must be at least as good as an explanation of observed facts about the concrete realm as T itself, since T adds to T* only in entailing the existence of causally inert entities like numbers. But anti-nominalists are not going to be per- suaded by such an argument: they will simply deny that there is any connection between causation and explanatory goodness of the kind the argument needs.24 Of course, such argumentative deadlocks are the norm across philosophy. But there is something very attractive about the idea that we should try to make progress in phi- losophy by learning from the disciplines in which progress is most manifest, namely the sciences. More specifi cally, the proposal is this: in the quest for a theory of good explanatory inference, we should take as our starting point the large and impressive body of case-by-case epistemological judgments shared by all scientifi c realists. We then decide what we ought to believe about controversial philosophical questions in accordance with whichever epistemological theory does the best job of accounting for and systematizing these data.25 In this methodological context, at least, the argument by analogy for the badness of T* is strong enough to be worth taking seriously. Even if we were to concede that T* was a bad theory, the debate would by no means be over. Nominalists could still look for alternatives to standard mathematics- laden scientifi c theories that explain the same data, but don’t entail that there are numbers, and are not similar to theories like T † in any way that would let an argu- ment by analogy get off the ground. This is the strategy favored by Hartry Field. In his book Science Without Numbers (1980), Field works out an elegant and clearly non-parasitic version of Newtonian gravitational theory, which entails the existence of nothing besides particles, space-time points, and space-time regions. He shows that, in conjunction with appropriate mathematical axioms and defi nitions of mixed math- ematico-physical predicates, this theory entails the “platonistic” theory it is meant to replace.26 However, Field’s program has yet to be carried out for theories like general relativity and quantum mechanics. At this stage, it is simply too early to say whether it is possible to fi nd nominalistic theories of these matters that are as free as Field’s theory is from any taint of similarity to bad theories like T †.27 So there is a strong motivation for those who think that we can, even now, reasonably believe nominalism – as opposed to merely adopting it as a working hypothesis – to resist the indispens- ability argument at an earlier stage, by maintaining that T* is a perfectly good theory as it stands, in spite of the analogies between it and T †. Cian Dorr 40 It is not hard to come up with differences between T* and T † that look like they might be epistemologically signifi cant. For example, there is the aforementioned fact that the additional strength of T over T* derives entirely from the postulation of enti- ties that are causally inert. But if we adhere to the “naturalistic” methodology for resolving such questions, we have to do more than merely point to one of these differences and claim it to be relevant. We will need to argue for its relevance by appealing to epistemological judgments common to all scientifi c realists, nominalists and anti-nominalists alike. Here’s what I think is the relevant difference. While both T* and T † result from the application of some complex modal operator to the original theory T, the opera- tors in question are of different logical kinds. The one in T † – ‘The facts about the positions and motions of atoms are consistent with the hypothesis that . . .’ – is, in essence, a possibility-operator. In the often helpful idiom of possible worlds, T † can be thought of as saying that there is some T-world where the facts about the positions and motions of atoms are just as they actually are. By contrast, the operator in T* – ‘If such-and-such mathematical axioms were true, it would be the case that . . .’ – has the logical properties of a necessity-operator. In possible-worlds terms, it says, in effect, that every world where the mathematical axioms are true that is like the actual world in concrete respects is a T-world.28 Now to the argument that this is an epistemologically important difference. Surely, if science can tell us about the unobservable at all, one thing it has told us is this: matter is not all alike, but comes in different kinds. Two bits of matter that are alike in respect of shape, size, and motion can still fail to be exactly alike – can fail to be duplicates. For example, they might fail to be duplicates by having different quantities or distributions of mass or charge, or by being composed of different kinds of elemen- tary particles. We have good reason to believe this in spite of the fact that for any theory T that entails that matter comes in different kinds, we can fi nd a weaker theory that makes the same predictions as T about the distribution of matter in space and time, without ruling out the hypothesis that matter is all alike.29 One way to formulate such a theory is to emulate T †, by saying something like ‘The facts about the distribu- tion of matter in space and time are consistent with the hypothesis that T.’ But pro- vided that T already contains a little mathematics – enough to entail the existence of sets of particles – there is no need to use the modal notion of consistency: we can achieve the same effect using existential quantifi ers. Suppose for the sake of concrete- ness that T entails that matter is not all alike by virtue of entailing that elementary particles come in different kinds: electrons, photons, quarks, neutrinos. . . . To formu- late a new theory T − consistent with the claim that matter is all alike, we will need to introduce some new variables v1, v2, v3, v4, . . . , corresponding to T’s one-place predicates ‘electron’, ‘photon’, ‘quark’, ‘neutrino’. . . . We then proceed in two steps. First, for each one-place atomic formula, like ‘x is an electron’ or ‘x is a photon’, in
T, we substitute the two-place formula ‘x ∈ vi’, where vi is the variable corresponding to the predicate in the original formula. Second, we insert existential quantifi ers at the beginning of the resulting formula, to bind all the new variables. The resulting theory T − says in effect that there is some way to assign sets of particles to be the extensions of the predicates ‘electron’, ‘photon’, ‘quark’, ‘neutrino’ . . . in such a way that T becomes true. This is clearly consistent with the hypothesis that all matter is There Are No Abstract Objects 41 alike, while having all the same consequences as T as regards the distribution of par- ticles across space and time.30 Since the existence of observationally adequate theories like T − in fact does nothing to undermine our reason to believe that matter comes in different kinds, T − must be a much worse theory than T. There is a general pattern here. If we want to weaken a theory so as to eliminate its commitment to some sort of hidden structure, we can very often do so by replacing the vocabulary purporting to characterize this structure with variables of an appropriate sort, bound by initial existential quantifi ers.31 So we can draw a general moral: since commitment to hidden structures often plays an essential part in good explanations, such existential quantifi cation must be a source of theoretical badness. When we replace a constant with a variable bound by an initial existential quantifi er, the resulting theory will typically be considerably worse than the one we started out with. Universal quantifi cation doesn’t seem to be a source of badness in the same sort of way. Indeed, when it is possible to weaken a theory by replacing one of its con- stants with a variable bound by an initial (restricted) universal quantifi er, the result of doing so is often considerably better than the original theory. Consider for example physical theories formulated in coordinate terms. When we assign ‘x’, ‘y’, and ‘z’ coordinates to particles, we certainly don’t mean to suggest that there is a single dis- tinguished, physically privileged coordinate system, concerning which it would make sense to wonder how far we are from the nearest axis. Rather, we are implicitly claim- ing that such-and-such equations hold in every admissible coordinate system. Similarly, when a theory uses numbers to measure mass or charge, it is often under- stood that the choice of a scale is arbitrary, so that what’s really being said is that the theory holds true for any admissible assignment of numbers.32 All this seems quite unproblematic. We would make these theories worse, not better, if we eliminated the implicit universal quantifi cation by positing a metaphysically special One True Coor- dinate System or Privileged Unit of Charge. The bad theories T † and T − seem to be bad in the very same way. So our explana- tion of the badness of T † should appeal to some factor common to T † and T −. What could this factor be? If T doesn’t contain any existential quantifi ers of the problematic kind, neither does T †. But it does contain the possibility-operator ‘The facts about the positions and motions of atoms are consistent with the hypothesis that. . . ’. And there is a notable logical affi nity between possibility-operators and existential quantifi ers, one that we exploit when we express claims about possibility in terms of possible worlds. (One need not endorse the unpopular opinion that ‘possibly’ means ‘in some possible world’ to see this.) Since this is the clearest point of similarity between T † and T − that does not also hold between T † and unproblematic theories involving universal quantifi cation over coordinate systems and the like, we can plausibly conclude that it is the basis for the epistemological similarity between the theories. Applying a possi- bility-operator to a theory, just like changing one of its constants to a variable bound by an initial existential quantifi er, generally leaves one with a much worse theory – a theory that, if true, would cry out for further explanation. If so, the example of T † gives us no reason to think that applying a necessity-operator to a theory, as we do when we move from T to T*, need do anything to make it worse. As far as the explana- tory status of T* is concerned, the analogy that turns out to be most important is not Cian Dorr 42 the one between T* and the bad theory T †, but the one between T* and the good theo- ries with initial universal quantifi ers ranging over coordinate systems. These are, I admit, dialectical baby steps. It’s all very well to point out suggestive analogies and disanalogies; the ultimate test for which analogies matter must come when we actually attempt to formulate general epistemological principles and see if we can get them to fi t with our intuitions about particular cases. I hope to take some steps in this direction in future work. But I hope that I have already said enough to shift the burden of proof back to those who maintain that belief in the reality of abstract objects can be justifi ed by their role in scientifi c explanations.
4 Abstract Objects in Philosophical Explanations
Much of philosophy is concerned with questions of the form ‘What is it to be F?’ or ‘What is it for something to R something else?’ – requests for analyses, or real defi ni- tions. Believers in abstract objects have frequently invoked them in their answers to such questions. They have given analyses on which apparently innocuous claims such as the following turn out to be implicitly about abstract objects:
(9) (a) Necessarily, all dogs are dogs. (b) Some people believe that penguins eat fi sh. (c) If I had missed the bus this morning I would have been late for class.
For example, (9a) is often held to be analyzed as ‘the proposition that all dogs are dogs is necessary’ (Bealer 1993); (9b) as ‘some people believe the proposition that penguins eat fi sh’ (Schiffer 2003: ch. 1); (9c) as ‘the closest worlds where I miss the bus are worlds in which I am late for class’ (Stalnaker 1968; Lewis 1973). Faced with analyses of this sort, nominalists have three options. First, they can accept the analyses, and conclude that the sentences in question are not true, taken in the fundamental sense. Second, they can suggest alternative analyses. Or, third, they can maintain that the relevant notions are primitive and unanalyzable. Taking the second option need not always mean a lot of hard philosophical work. A lazy nominalist can simply take an analysis mentioning abstract objects, and insert something like, ‘If there were abstract objects, it would be the case that . . .’ at the beginning. The hardest cases for nominalism are those where the fi rst of these options, and the lazy version of the second option, can be ruled out. For example, if we want to use counterfactuals in paraphrasing ordinary abstract-object talk, as suggested in section 2, we obviously can’t hold that sentences involving the counterfactual condi- tional are never true in the fundamental sense. And it would be circular to analyze ‘If it were the case that P, it would be the case that Q’ in general as ‘If there were abstract objects, the closest possible worlds in which P would be worlds in which Q.’ So we are left with a choice between taking counterfactual conditionals as primitive and maintaining they can be analyzed without bringing in abstract objects. Despite its importance, I won’t have anything more to say here about the diffi cult task of providing an nominalistically acceptable account of counterfactuals.33 Instead, There Are No Abstract Objects 43 I will consider another hard case, that of basic physical predicates. What is it for something to be an electron, for example? Some anti-nominalists, such as David Armstrong (1978a, 1978b), answer as follows:
(10) To be an electron is to instantiate the property electronhood.
Surely the fi rst option – accept the analysis, and concede that there aren’t, funda- mentally speaking, any electrons – is not a real option here. (Not that it is so obvious that there are electrons, fundamentally speaking – but if there aren’t, it is not because there are no properties!) Likewise, the lazy version of the second option – claiming that to be an electron is to be something which would instantiate electronhood if there were abstract objects – is deeply unattractive. Surely, if this counterfactual is true of an object, it is true because it is an electron, and not the other way round. So we are left with the choice of taking ‘electron’ as primitive, or fi nding some analysis quite different in character from (10). For too long, analyses like (10) were considered only in the context of the claim that all instances of the following schemata are true, no matter what predicates we substitute for ‘F’ and ‘R’:
(11) To be F is to instantiate the property, F-ness. (12) For x to R y is for x to bear the relation, R-ing, to y.
But there are at least two good reasons to reject this sweeping attempt to provide real defi nitions for all predicates at once. First, there is Bradley’s regress (Bradley 1897: ch.2). If we accept the following instance of (12)
(13) For x to instantiate y is for x to bear the relation, instantiation, to y we have taken the fi rst step in a vicious infi nite regress. We take the next step when we accept the corresponding analysis of the predicate on the right-hand side of (13):
(14) For x to bear z to y is for x, y, and z to stand in the relation, bearing.
And so on. This can’t go on forever: analysis must come to an end somewhere. Second, many predicates already have real defi nitions not in the form of (11) or (12). For example, we seem to have learned by doing science that
(15) To be a helium atom is to be an atom whose nucleus contains exactly two protons.
If we combined this with the relevant instance of (11)
(16) To be a helium atom is to instantiate the property, being a helium atom we could presumably conclude that
(17) To instantiate the property, being a helium atom, is to be an atom whose nucleus contains exactly two protons. Cian Dorr 44 I can see how (17) might be true if we were thinking of the property being a helium atom as existing only in a superfi cial sense, so that truths that appear on the surface to involve it turn out, on deeper analysis, to be truths of a different sort. But how could (17) be true if the property is a fully real thing, existing in the fundamental sense? I can’t see how an account of what it is for an object to instantiate a certain real entity could fail to mention that entity at all, any more than an account of what it is for a region of space to contain a certain real particle could fail to mention that particle at all.34 So much the worse for (11) and (12) in full generality. But, as Armstrong forcefully argues, it would be a mistake to dismiss the idea that some instances of these schemata are true. According to Armstrong, (11) holds only for an elite minority of predicates, which it is up to physics to identify.35 It is a good question what, if anything, a nominalist should put in place of these analyses. One might reasonably object that the question what it is to be an electron is one we should not expect to be able to answer from the armchair. Physicists discovered that to be a helium atom is to be an atom whose nucleus contains exactly two protons, and that to be a proton is to be a complex of quarks of certain kinds bound together in certain ways; it would be foolish for metaphysicians to rule out the possibility of some similar discovery about what it is to be an electron.36 But there is still a chal- lenge: the analysis of “physical” predicates like ‘helium atom’, ‘proton’, and ‘electron’ in terms of other such predicates cannot go on forever. Eventually (since there can’t be infi nitely regressive or circular real defi nitions) there must be a basic physical predicate that doesn’t have a real defi nition involving other physical predicates – either because it has no defi nition at all, or because it has a defi nition that doesn’t involve any physical predicates, e.g. a defi nition of the form of (11). The real chal- lenge for the nominalist is to say something about what happens then. From now on I’ll assume for ease of exposition that ‘electron’ is such a basic physical predicate. So, why would one think it better to accept (10) rather than taking ‘electron’ as primitive, or analyzing it in some other way? The appeal of (10) becomes apparent only when we turn to another class of predicates for which analyses in terms of abstract objects are often proposed, namely predicates having to do with resemblance. A good example to focus on is the notion of duplication, or perfect intrinsic resem- blance.37 Suppose that the believer in properties analyzes this as follows:
(18) For x to be a duplicate of y is for x and y to instantiate exactly the same things.38
Putting this together with (10), we can account for the necessity of the following claim:
(D) Whenever x is an electron and y is a duplicate of x, y is an electron.
For, when we replace the predicates ‘electron’ and ‘duplicate’ with the analysis given by (10) and (18), (D) reduces to an elementary logical truth:
(D*) Whenever x instantiates electronhood and y and x instantiate exactly the same things, y instantiates electronhood. There Are No Abstract Objects 45 This is a signifi cant explanatory achievement. It is not obvious what nominalists should put in its place. There are three basic options, short of simply denying the necessity of (D).39 First, we could maintain that the necessity of (D) follows from an analysis of ‘duplicate’ in which the predicate ‘electron’ occurs, presumably along with other predicates. Second, we could maintain that it follows from an analysis of ‘electron’, either alone or in conjunction with some analysis of ‘duplicate’. Or, third, we could maintain that the necessity of (D) cannot be explained in terms of the real defi nitions of its constituent predicates. I will consider each of these strategies in turn.
The physical strategy Although there have been few explicit defenses of the fi rst strategy, I suspect that it is in fact quite popular.40 The idea is straightforward: just as we looked to science to provide us with real defi nitions of predicates like ‘water’ and ‘helium atom’, so we must ultimately look to science to provide us with real defi nitions of ‘duplicate’ and other predicates of the same general, structural sort. What might such an analysis look like? Bracketing complications raised by duplication amongst complex objects, all we need is a simple conjunction of biconditionals, one for each monadic basic physical predicate:
(19) For x to be a duplicate of y is for it to be the case that x is an electron iff y is an electron, and x is a quark iff y is a quark, and . . .41
Since this obviously suffi ces to entail the necessity of (D), it leaves the nominalist free to take ‘electron’ as primitive. The main problem I see for the physical strategy is its inability to accommodate what I will call the Alien Properties Intuition (cf. Lewis 1983: 158ff.). Roughly speak- ing, this is the intuition that that there might be alien properties, not identical to or constructed from any actual properties.42 Of course, nominalists can’t accept that way of putting it: but we can still think that when believers in properties talk about situ- ations containing alien properties, they are talking about genuine possibilities, even if they are misdescribing them. Typically, these possibilities will contain things that fail to be duplicates even though they are indiscernible as far as the basic predicates of actual-world physics are concerned:
(20) Possibly, there are two simple things x and y that are not duplicates, although x is an electron iff y is an electron, and x is a quark iff y is a quark, and . . .
One special case concerns the possibility that nothing falls under any of the basic predicates of actual-world physics. The proponent of the physical strategy is com- mitted to the claim that if this were the case, the world would be completely homo- geneous. Any two simple things would be duplicates, and indeed qualitatively indiscernible. But this seems wrong. Intuitively, there could be all kinds of interesting goings-on at possible worlds where nothing falls under any of the basic predicates of actual-world physics: they could be every bit as richly varied in their own ways as the actual world is in its.43 Cian Dorr 46 If this is right, it shows that ‘duplicate’ is just not the sort of predicate that could have an analysis in physical terms. (Arguments of the same sort will apply to other predicates with a similarly “structural” or “topic-neutral” character.) That’s a lot to rest on intuitions about such remote possibilities. Although I feel the pull of the Alien Properties Intuition quite strongly, I’m not sure that its force would be strong enough to make me reject nominalism, if it turned out that the only way to continue being a nominalist was to give it up. But let’s press on and consider the other options.
The structural strategy The second option for the nominalist is to explain the necessity of (D) as following from a real defi nition of ‘electron’, perhaps in conjunction with a real defi nition of ‘duplicate’. What might the analysis of a basic physical predicate like ‘electron’ look like? In the fi rst half of the twentieth century, it was widely believed that physical predicates could be analyzed in terms of “observational” predicates: to be an electron is to be something disposed to make white tracks of a certain shape in cloud chambers of a certain design, etc. This sort of view is now deservedly unpopular, for reasons I won’t go into. Even if something like this were true, it wouldn’t really help us, since observa- tional predicates like ‘white’ raise the same essential problem for the nominalist as basic physical predicates like ‘electron’. Just as any duplicate of an electron must be an elec- tron, so any duplicate of a white object must be white.44 And as far as the explanation of this necessity is concerned, the analogue of the physical strategy – an analysis of ‘duplicate’ in terms of observational predicates like ‘white’ – looks even more problem- atic than the corresponding explanation of (D). The fact that two objects cannot be distinguished by our observational predicates doesn’t entail that they are duplicates. How could we analyze ‘electron’ without merely postponing the problem in this way? We might, for example, be Resemblance Nominalists, and attempt – somehow or other – to analyze ‘electron’ (along with everything else) entirely in terms of resemblance. That need not mean that the only non-logical predicate we can use in our analysis is ‘x resembles y’: most Resemblance Nominalists have allowed themselves more fl exible primitives to work with. For example, Price (1953) seems to treat claims of the form ‘x1 resembles x2 at least as much as y1 resembles y2’ as primitive. It is not hard to see how one might go about analyzing ‘duplicate’ in terms of predicates like these. One could say, for example, that for x to be a duplicate of y is for x to resemble exactly the same things as y, or for x to resemble y at least as much as y resembles itself. While we might object to these analyses on other grounds, at least there is no confl ict with the Alien Properties Intuition. There are various other notions with the same abstract, structural character as resemblance – notions which would not look out of place in an analysis of ‘duplicate’, and hence could occur in an analysis of ‘electron’ without merely postponing our problem. For example, Natural Class Nominalists take as primitive the notion of a “natural” class – informally speaking, a class of things that all resemble one another in some one respect and resemble nothing else in that respect.45 Taken at face value, this predicate is of no use to a nominalist in my sense, who denies that there really are any classes (sets). But this is one of those cases where a claim about classes can be regarded as a misleading way of saying something properly expressed using plurals. There Are No Abstract Objects 47 That is, instead of saying that the class of electrons (singular) is natural, we should really say that the electrons (plural) are, collectively, natural.46 Again, it is not hard to see how this notion might feature in a plausible analysis of ‘duplicate’. For example, we could say that for x to be a duplicate of y is for it to be the case that whenever some things are natural, x is one of them iff y is. Mereological predicates like ‘part of’ seem to belong in the same “structural” cate- gory.47 Other cases are harder to adjudicate. Perhaps certain causal vocabulary should be allowed. Some (e.g. Campbell 1991) even put spatiotemporal predicates in this category – though from my point of view, the inclusion of ‘before’ or ‘between’ in a defi nition of duplication looks only slightly less problematic than the inclusion of ‘electron’ or ‘quark’. But we don’t really need to decide these questions: the hard part is seeing how ‘electron’ could have a defi nition in terms of any of these materials. It has generally been assumed in discussions of Resemblance Nominalism that the analyses of most ordinary predicates will involve reference to certain “paradigm” particulars. For example, adapting a suggestion of Price’s (1953), one might propose the following analysis for ‘electron’:
(21) To be an electron is to resemble each of e1, e2, . . . en at least as much as any two of them resemble one another.
(There are various possible complications: for example, one might want to have ‘anti-paradigms’ which one must fail to resemble to be an electron.48) One initially off-putting fact about analyses like (21) is the arbitrariness of the choice of paradigms. It is not plausible that we are simply ignorant of the identity of the paradigm electrons. But this doesn’t seem like a serious problem, since it is open to the proponent of the strategy to claim that there is no determinate fact of the matter as regards the identity of the paradigms: in a sense, ‘electron’ is vague, though we may know a priori that it has no actual borderline cases.49 A second objection to analyses like (21) makes use of the notion of intrinsicness: whether something is an electron is a matter of what that thing is like intrinsically; whether something resembles the paradigm electrons is not; hence, it can’t be the case that to be an electron just is to resemble the paradigm electrons.50 How should a proponent of the structural strategy understand the notion of intrinsicness that features in this argument? On the one hand, an intrinsic characterization of something is supposed to be one that neither explicitly nor implicitly refers to, or quantifi es over, anything apart from the thing in question and its parts (and perhaps also the proper- ties and relations they instantiate, and the sets they are members of – but these are things a nominalist doesn’t believe in). On this demanding conception, adopting the structural strategy requires one to say that only a very few predicates, like ‘resembles itself’ or ‘has exactly seven duplicate parts’, characterize things intrinsically. Clearly, if ‘electron’ has an analysis in “structural” terms, it will not count as intrinsic in this demanding sense. On the other hand, it is also supposed to be necessary that any two things that have exactly the same intrinsic features are duplicates. But this clearly won’t be the case if we adopt the demanding conception. If any form of the structural strategy is correct, it is pretty likely that any two simple things are alike in all the respects that count as intrinsic on the demanding conception, whether or not they Cian Dorr 48 are duplicates. Given these confl icting desiderata, perhaps the best thing to say is that ‘intrinsic’ is ambiguous between the demanding sense and some more liberal, but more metaphysically arbitrary, sense on which basic physical predicates do count as characterizing things intrinsically. But whatever we end up saying, we must concede that the structural strategy will be unacceptable to those who think they have a fi rm grip on a notion of intrinsicness satisfying both desiderata. There is a third problem for the strategy of mentioning particular objects in the defi nition of ‘electron’, which seems to me by far the most serious. Namely: any workable analysis of this sort will entail that it is necessary that at least some of those objects exist, if there are any electrons (Armstrong 1978a: 51–3; Van Cleve 1994: 579). This is seriously implausible. Surely there could still have been electrons even if no actual particulars had existed. Let’s call this the Alien Particulars Intuition. If we accept it, the idea of using paradigms in the analysis of basic physical predicates must be given up.51 The Alien Particulars Intuition rules out just about any structural analysis of ‘elec- tron’ involving reference to particular objects. That leaves us with purely general structural analyses: analyses according to which the facts about the structure of the world are already suffi cient to fi x which things are electrons, irrespective of which particular objects occupy the positions in the structure. While this approach is consistent with the Alien Particulars Intuition, it has its own diffi culties with modal intuition. Many philosophers (e.g. Lewis 1986, Armstrong 1989a) accept what I will call the Humean Intuitions: they think that there are few, if any, interesting structural conditions something must satisfy in order to fall under a basic physical predicate. Instead of obeying the actual laws of nature, electrons could be distributed in some very different way. For example, there could be exactly seven electrons, evenly spaced in a straight line and motionless in otherwise empty space. Likewise, there could be exactly seven non-electrons, evenly spaced in a straight line and motionless in otherwise empty space. But if both of these situations are pos- sible, it cannot be the case that whether something is an electron depends only on its place in the structure of resemblances, natural classes, or whatever: these kinds of structural facts are exactly the same at the world with seven electrons and at the world with seven non-electrons. But the Humean Intuitions are by no means uncontroversial. Several philosophers have endorsed, for reasons that don’t on their face have much to do with nominalism, the “dispositional essentialist” view that it is necessary that the objects falling under a basic physical predicate like ‘electron’ should play a certain characteristic role in the laws of nature.52 On this sort of view, many of the truths that would traditionally be classifi ed as laws of nature – “nomological necessities” – will in fact be metaphysi- cally necessary. For example, it might be necessary that if there are any electrons, they repel one another and attract any protons there might be.53 If we want to maintain that basic physical predicates have structural defi nitions, can we at least avoid having to classify as metaphysically necessary any truths we would not otherwise have any reason to classify even as nomologically necessary? Unfortunately the answer seems to be ‘no’, at least if we hold onto our assumption that ‘electron’ is a basic physical predicate.54 The problem is that electrons and posi- trons play symmetric roles in current physical theories. If one reinterprets ‘electron’ There Are No Abstract Objects 49 as standing for positrons, and ‘positron’ as standing for electrons, while making certain other substitutions of a similar nature, the theory will still be true on the new interpretation. This means that for a purely general structural defi nition of ‘electron’ to avoid incorrectly classifying positrons as electrons, it will have to appeal to some distinguishing characteristic of the electrons that a Humean would not even regard as nomologically necessary. For example, if there are in fact many more electrons than positrons, we might say that to be an electron is to be a member of the largest natural class playing some characteristic nomological role. If the predominance of electrons over positrons is a merely local phenomenon, we will have to rely on more subtle differences – e.g. the fact that electrons outnumber positrons in the region of space that surrounds a planet that is appropriately similar in structural respects to the planet Earth as it actually is.55 Clearly it would be unrealistic to expect precision here. If the usual laws of nature don’t suffi ce to do the job of distinguishing electrons from positrons, we’re going to have to settle for considerable vagueness as regards just how much of the actual “electron role” some things would have to play to count as electrons. This is certainly a bit unsettling. But I am inclined to think that if we can reconcile ourselves to giving up the Humean Intuitions, we should not be too concerned to fi nd that some of the truths we end up counting as metaphysically necessary are truths that someone in the grip of those intuitions wouldn’t even think of as laws of nature. When the dispositional essentialist idea that objects at a world very unlike the actual world (such as the world with seven static identical particles) could not be electrons starts to seem compelling, it does so precisely because one cannot see what could make objects at a world like that count as electrons. Once we start expecting non-trivial answers to questions of this sort, they will seem equally pressing when we consider the putative possibility of electrons and positrons switching roles. We will want to know what could make these things count as the positrons and these as the electrons, rather than vice versa. For a proponent of the Humean Intuitions, this is of course a bad question. But in my experience it is not so hard to get people to feel its force.56
Brute necessities If all this is right, the only way to explain necessities like (D) without giving up the Alien Properties Intuition, the Alien Particulars Intuition, or the Humean Intuitions is to reject nominalism. Before we can evaluate this argument, we must ask why such necessities should need explanation. What if a nominalist were to maintain that they are “brute” necessities, which cannot be explained by any analyses of their constituent predicates? One might object to this view on the grounds that necessities never are brute. The only genuinely necessary truths – one might maintain – are those that reduce, upon analysis, to truths of logic (in some narrowly delimited sense of ‘logic’). This is the strongest form a ban on brute necessities might take: various natural weakenings would still work in the argument. I will briefl y mention three. (1) If we wanted to allow for facts about the essences of objects as a distinct source of neces- sity (see Fine 1994), we might still hold that all purely general necessary truths – all truths that do not either implicitly or explicitly involve reference to any specifi c entities – reduce to truths of logic. (2) If we were persuaded by Kripke’s argument Cian Dorr 50 (1972: 156) for the necessity of sentences like ‘there are no unicorns’ and ‘there is no phlogiston’ but despaired of fi nding appropriate analyses of ‘unicorn’ and ‘phlogiston’, we might want to make a special exception for such “semantically defective” predi- cates.57 (3) More ambitiously, we might want to make an exception for a category of “non-factual” vocabulary (perhaps including evaluative terms like ‘good’) whose com- municative function is not, strictly speaking, that of expressing our beliefs about reality, but something quite different.58 Since it is impossible to analyze the non- factual in terms of the factual, most necessary truths essentially involving non-factual vocabulary (e.g., ‘love for one’s children is good’) will not admit of reduction to logical truths. None of these weakenings is of much use to the nominalist faced with the task of explaining the necessity of (D). (1) If we mentioned some particular objects in the analysis of ‘electron’, we might in principle attempt to explain the necessity of (D) in terms of the essences of these objects; but the Alien Particulars Intuition rules out such analyses. (2) Only an idealist could take seriously the idea that all basic physical predicates are semantically defective; and the idealist would face an exactly parallel problem involving mental predicates. And (3) if the distinction between non-factual and factual vocabulary makes sense at all, ‘electron’ and ‘duplicate’ seem to belong on the factual side of the line. But orthodox anti-nominalists are in no position to accept the claim that there are no brute necessities, even when these qualifi cations are taken into account. For most anti-nominalists will be committed to the necessity of a wide range of axioms specify- ing conditions under which numbers, sets, and logically complex properties and rela- tions exist. Since these axioms are existential in form, there is no hope of reducing them to truths of standard logic just by analyzing predicates like ‘set’ and ‘property’. More- over, if they are true, many of these axioms are purely general, semantically non-defec- tive, and fully factual. Some anti-nominalists (e.g. Bealer 1982) advocate a wider use of ‘logic’ on which these axioms will count as logical truths. But I can see no indepen- dent motivation for a principle that would allow such axioms to be necessary while requiring the necessity of (D) to be explained by analyses of ‘duplicate’ and ‘electron’. In the absence of such a principle, presumably the debate will have to turn on considerations of economy. The question will be which of the available theories makes do with the smallest, simplest set of brute necessities.59 But it is not at all clear that anti-nominalists will do better by this criterion. And even if it were, it would not be clear how this kind of economy should be weighed against the considerations of ontological economy that favor the nominalist. On the whole, I doubt nominalists have much to fear once the dispute turns into a contest of economy. However, not all anti-nominalists accept the necessity of all the usual axioms. For example, some of them have denied that it is necessary that for any two properties p and q, there is (in the fundamental sense) a property instantiated by exactly those objects that instantiate either p or q.60 We can imagine an extreme version of this view, on which the only necessary general principles about properties and instantia- tion are those that reduce, under analysis, to truths of logic (narrowly conceived). A proponent of this view would have a principled reason for insisting that the necessity of (D) must be explained by analyzing ‘electron’ or ‘duplicate’. But the prospects for this kind of view seem poor. Consider, fi rst, a version of the view according to which the predicate ‘instantiates’ is primitive and unanalyzable, so There Are No Abstract Objects 51 that for sentences whose only piece of non-logical vocabulary is the predicate ‘instantiates’, metaphysical necessity coincides with strictly logical necessity, and metaphysical possibility with strictly logical possibility. This seems utterly incredible. Not just any old collection of points and arrows connecting them represent a way things might be, if the arrow is interpreted as meaning ‘instantiates’. Is there a possible world with just 17 entities, a1 . . . a17, such that a1 instantiates a2, a2 instantiates
a3 . . . and a17 instantiates a1? Is there a possible world just like the actual world in respect of the structure of instantiation except that my nose (or its counterpart) instantiates my little fi nger? Is there a possible world structurally like the actual world except that any two things instantiate one another if they are actually less than a mile apart? Surely not. Hence, an anti-nominalist who wants to uphold the ban on brute necessities will have to fi nd some analysis of ‘instantiates’ that can account for these impossibilities. But what might such an analysis look like? I can’t see how anything simple or intui- tive could do the job. The only way I can see to get the right necessary truths to fall out of a defi nition of ‘instantiates’ (without going so far as to make it impossible for anything ever to instantiate anything else) would be to tailor the defi nition in such a way that exotic situations like those discussed in the previous paragraph get reclassifi ed, in effect, as possibilities in which nothing instantiates anything else. For example, one might begin with a primitive notion of “proto-instantiation,” and propose that for x to instantiate y is for x to proto-instantiate y at a world where the facts about proto-instantiation make up the right kind of overall pattern.61 This feels like cheating to me. If we give ourselves free rein to make up new predicates, conve- niently free from any involvement in our pre-existing modal beliefs, it will become a trivial task to fi nd analyses of our old predicates on which all the sentences that intuitively strike us as necessary, and none of the sentences that intuitively strike us as contingent, can be reduced to logical truths.62 We simply analyze each predicate ‘F’ as ‘proto-F such that P’, where P is the conjunction of all the sentences we want to turn out to be necessary, with ‘proto-’ inserted in front of each predicate.63 Unfor- tunately, I am not at all sure how to turn this feeling into an argument against the proposal. I am tempted to protest that the new predicate ‘proto-instantiates’ is simply unintelligible. But this kind of objection needs to be handled very delicately if it is not to rule out legitimate conceptual innovation. Another tempting line of argument is epistemological: why should we have any confi dence that the world has the distinc- tive kind of structure it would need to have to contain instantiation rather than mere proto-instantiation?64 But anyone who wields localized skeptical arguments like this one must be prepared for the inevitable response: ‘You tell me how you know that you’re not a brain in a vat, and I’ll tell you how I know this fact you allege I could not know if my account of its nature were correct.’ In any case, it is clear that if we do allow “cheating” analyses like this, nominalists have nothing to worry about from the demand for an explanation of necessities like (D), even if they are persuaded by the objections to the physical strategy and the structural strategy discussed above. They can use the very same trick, e.g. by analyzing ‘electron’ as ‘proto-electron that is such that any duplicate of a proto-electron is a proto-electron’, or by analyzing ‘x is a duplicate of y’ as ‘x is a proto-duplicate of y, and any proto-duplicate of an electron is an electron’. Cian Dorr 52 To sum up: we have not managed to fi nd a stable argument against nominalism based on the use of abstract objects in explaining necessities like (D). Instead, we have stumbled on an argument in favor of nominalism. If there are, in the fundamental sense, numbers, properties, relations, or sets, then there are necessary truths about these things that cannot (assuming we can somehow rule out the trivializing ‘proto- instantiation’ move) be reduced to truths of logic. Thus, only the nominalist, who denies that there are any such things, can adequately respect the idea that there are no brute necessities.65 I think this is quite a powerful argument. To anyone not antecedently convinced of the falsity of nominalism, the idea of a metaphysically necessary truth whose necessity does not fl ow from real defi nitions plus logic really should seem quite strange. A notion of necessity that allowed for such necessary truths would seem uncomfortably like nothing more than an extra-strong variety of nomological necessity. But when some- thing strikes us as impossible – say, the hypothesis that some duplicate of an electron is not itself an electron – we don’t just think of it as ruled out by a “law of metaphys- ics”: we feel that in some important sense, the idea just makes no sense at all. The notion of necessity involved in such intuitions is absolute: it is something that could not be strengthened any further without changing its character in some fundamental way. It is hard to see how any notion of necessity weaker than the notion of reducibility to logical truth could be absolute in this way.66 Of course, even for a nominalist, the task of reconciling the ban on brute necessi- ties with the facts about necessity and possibility is not an easy one (assuming we can somehow rule out the trivializing ‘proto-electron’ move). Resemblance Nominal- ism, for example, is not an option. All the resemblance-predicates I can think of are involved in manifestly necessary truths that are not logical truths, and contain no other non-logical vocabulary. (A simple example: the necessary truth that any dupli- cate of a duplicate of an object is a duplicate of that object.) If they are not brute, these necessities must be explained by analysing resemblance-predicates in terms of predicates of some other sort. But the outlook for other nominalist programmes is better. The physical strategy provides one possible approach for foes of brute neces- sities: it is at least not obvious that there are any non-logical necessary truths involv- ing only basic physical predicates.67 And although Resemblance Nominalism is ruled out, Natural Class Nominalism remains a live option for those who prefer the structural strategy. For there is no clear reason why settling the question whether some given objects are (collectively) natural should entail anything at all about the naturalness of any other objects. Thus, for those who want to uphold a version of the ban on brute necessities strong enough to place a serious constraint on the shape of our metaphysical theorizing, nominalism provides at least two promising research pro- grams, while anti-nominalism provides none.68
Notes
1 The latter example is due to van Inwagen (2004). 2 Similarly, our ordinary use of number-talk commits us to treating the argument ‘Either there are no planets, or there is a planet; therefore, either the number of planets is zero, There Are No Abstract Objects 53 or the number of planets is at least one’ as valid. Since the premise of this argument is analytically true (true just in virtue of meaning), the conclusion must be too, on its ordi- nary use. But it cannot be analytic when taken in the fundamental sense: fundamentally speaking, it is not an analytic truth that there are numbers, since it is not an analytic truth that there is anything at all. As Hume and Kant maintained in criticizing the standard a priori arguments for the existence of God, denials of existence – when taken in the fundamental sense – can never be self-contradictory. 3 On defi ning ‘abstract’, see Lewis (1986: 82–6), Burgess and Rosen (1997: 13–25), and section 1 of Chris Swoyer’s companion chapter in this volume (1.1). 4 There is a complication here, which arises when we consider a sentence like
(*) ‘There are no numbers’ is true.
Surely, if there are (fundamentally speaking) no numbers, sets, properties, or relations, there also aren’t (fundamentally speaking) any sentences (i.e. sentence types). If so, it is hard to see how (*) could be true in the fundamental sense, given that the term in quotation marks purports to refer to a sentence. And there is certainly a superfi cial use of (*) on which it expresses a falsehood, just as (4*) expresses a truth. Nevertheless, it seems clear that there is some way of using (*) on which I should fi nd it acceptable, given my thesis. (For more details see the discussion in Yablo (2001) of ‘The number of numbers is zero.’) Indeed, when I make claims about sentences, theories, claims, beliefs, views, hypotheses, and so forth in the course of this chapter, I will often be talking in this way: still superfi cial, but closer to the fundamental than most ordinary uses of this sort of language. 5 Stanley (2001) and Burgess and Rosen (2005) argue on empirical grounds against this claim, which they attribute to Yablo (2000) – although Yablo’s (2001) response to Stanley suggests that he may not in fact accept it. 6 “Something from nothing” arguments along these lines have been defended by, among others, Alston (1963), Wright (1983), and Schiffer (2003). Many other arguments against nominalism seem to embed arguments of this kind at crucial points. For example, Russell (1912: ch. 10) gives a celebrated argument for the existence of relations, which depends on the premise that one thing cannot resemble another thing without bearing the relation of resemblance to it. 7 See also chapter 9 of this volume. 8 I make this attempt in Dorr (2005). 9 This is probably the best way to make sense of the idea that the existence of abstract objects ‘is just a matter of linguistic convention.’ On its more straightforward interpreta- tions, this slogan is deeply problematic. On one interpretation, it confl icts with the appar- ently obvious fact that if numbers exist, they would still have existed even if there had never been any human beings. On another interpretation, it is trivial, since every true sentence is true partly in virtue of meaning what it does, and every sentence means what it does partly in virtue of linguistic convention. On a third interpretation, it requires us to make sense of the notoriously diffi cult idea of a sentence being true wholly in virtue of our conventions. 10 The idea of “paraphrase” as a method for reducing one’s “ontological commitments” is introduced in Quine (1948). But a warning is in order: given Quine’s verifi cationism and his rejection of the notions of sameness and difference in meaning (1951), any apparent similarities between the project Quine calls ‘ontology’ and my project of investigating what there is in the fundamental sense are probably quite misleading. 11 These paraphrases will work only if the kind of superfi cial use of (6b) we are trying to capture is one on which it can be true only if planets exist in the fundamental sense. In Cian Dorr 54 my view, ordinary uses of sentences like (6b) and (6a) do not require planets to exist in the fundamental sense – they are consistent with the claim that, fundamentally speaking, there are no composite objects at all (see Dorr 2005). But in the context of the present discussion, we can harmlessly ignore divergences between the fundamental and superfi cial ways of talking that have nothing to do with abstract objects. 12 For arguments that an adequate system of nominalistic paraphrases of sentences about abstract objects is impossible, see Pap (1959–60), Jackson (1977), and van Inwagen (1977, 2004). 13 Two qualifi cations. First, we must distinguish the question whether an adequate system of paraphrases is possible in principle from the question whether we are in a position to provide such paraphrases. One can understand two languages without being in a position to translate between them – for example, a bilingual speaker of English and French might understand ‘elm’ and ‘orme’ without knowing that they refer to the same kind of tree – and this could be our situation as bilingual speakers of “fundamental English” and “superfi cial English.” Second: it would be enough for the paraphrases to be statable in a version of English supplemented with arbitrarily powerful devices of infi nitary conjunction, disjunc- tion, and quantifi cation. It could well be among the benefi ts of the superfi cial way of talking that it gives us a fi nite way of expressing what would otherwise require infi nitary resources. 14 For other attempts to paraphrase problematic sentences by embedding them in the scope of some sort of modal operator, see, e.g., Putnam (1967), van Fraassen (1980), Rosen (1990), and Yablo (2000). 15 It is a good question how a nominalist should understand the clause ‘. . . and the concrete world were just as it actually is.’ In section 4, I will consider a question that raises many of the same issues, namely how a nominalist should understand the notion of duplication, or exact intrinsic similarity. 16 Kripke (1972: 156) argues infl uentially for the claim that it is metaphysically impossible that there should be unicorns. 17 For those who still think that the fact that nominalism is necessary if true poses a problem, the literature on nominalism contains several proposals (e.g. Chihara 1990; Hellman 1989) that aim to get around this supposed problem by fi nding some metaphysically possible “surrogates” to play the role the impossible axioms about abstract objects play in my paraphrases. I see no problem of principle with these approaches, although they have the practical disadvantage that they can be used to write down paraphrases only for sentences in which all the predicates that apply to abstract objects are ones we already know how to analyze in terms of the predicates that appear in the axioms. 18 Even if there happen to be some “uninvolved” things that have no positive features incon- sistent with being a number, etc., it might still be the case that none of them are numbers. Some philosophers (e.g., Lewis 1993) have argued that the meaning of words like ‘number’ is very undemanding, so that the existence of suffi ciently many uninvolved objects is all that is required for some of them to count as “numbers.” The idea is that provided that there are enough uninvolved objects, names like ‘0’, ‘1’, ‘2’, . . . will refer indeterminately to all of them, so that there is no fact of the matter as regards which of them is 0, which is 1, etc. I see no sound motivation for such a prima facie surprising proposal. If we were determined to interpret ordinary number-talk as fundamental, there would indeed be pres- sure to interpret ‘number’ as undemandingly as possible so as to make it as easy as possible for ‘there are numbers’ to be true. But I have already argued against such an interpreta- tion: to do justice to our ordinary use, ‘there are numbers’ needs to be interpreted as analytic, and it is not analytic that there are infi nitely many uninvolved objects. On a more natural interpretation of ‘number’, the mere existence of infi nitely many uninvolved There Are No Abstract Objects 55 objects would not be enough for ‘there are numbers’ to be true in the fundamental sense: certain of these objects would, in addition, have to be structured in the distinctive way captured by the axioms of number theory. Understood in this way, it is clear that even if for some reason we were confi dent that there were lots of uninvolved objects, we should still, in the absence of positive argument, give little credence to the hypothesis that any of these objects are numbers. 19 I don’t mean these labels to suggest that either of these kinds of explanation is the exclu- sive province of philosophers or scientists. 20 I don’t mean to suggest that Burgess and Rosen intend to be arguing for that conclusion. It is clear from their most recent paper (2005) that, to the extent that they have any quarrel with what I call ‘nominalism’, it is because they think that not enough has been done to explain what it means to ask whether things of a certain kind exist in the fundamental sense. 21 At least in a superfi cial sense. The claim that subatomic particles exist in the fundamental sense seems much riskier, given the current state of our understanding of quantum fi eld theories. 22 Cf. Field (1988: 260). 23 As far as I know, the most explicit presentation of an argument by analogy along these lines in the literature is Field’s (1988: 260–1). However, Field’s argument deals in the fi rst instance, not with theories like T*, but with theories of the form ‘All the facts about con- crete objects are consistent with the hypothesis that T’, which are far more immediately analogous to T †. He conjectures that his conclusions about these theories generalize, so that every device by which one might “modalize away” the mathematical content of a theory is immediately analogous to some device by which one might “modalize away” its commitment to subatomic particles. But this conjecture seems to be false: there just isn’t any way for an eliminativist about subatomic particles to directly mimic the procedure used by the nominalist in formulating T*. For what would take the place of the mathemati- cal axioms in the antecedent of the conditional? The obvious candidate would be some claim to the effect that an atom is located at a point iff that point is the centre of mass of some subatomic particles bound together in such-and-such ways. But this won’t work. Given any world of atoms, there will be many different worlds where subatomic particles are added in such a way as to make such a claim true while leaving the distribution of atoms unchanged; T will be true at, at most, a few of these worlds. And it is hard to see what, other than the truth of T, could justify the claim that some T-world is closer to actuality than any non-T world. 24 And they are right that if there is any connection here, it is not a very straightforward one. Since we have reason to believe that the world is not about to end, some theories T that entail that it is not about to end must be explanatorily better, in the relevant sense, than any theory of the form ‘Up to now, everything has been just as if T’ – despite the fact that, since they are located in the future, the additional entities posited by T play no role in causing the observations that constitute our evidence for T. 25 This is my favorite candidate to be the referent of the coveted label ‘naturalism’ (Burgess and Rosen 1997: 33). 26 The result I have just mentioned holds only for the versions of Field’s theory that use logical resources stronger than fi rst-order logic. For debate about whether a commitment to the superiority of theories using fi rst-order logic might revive the indispensability argu- ment, see Shapiro (1983) and Field (1985). 27 And even if the program were fully successful in basic physics, there would be a diffi cult further question about the role of mathematics in “higher-level” sciences like statistical mechanics and population ecology. Cian Dorr 56 28 Note that if the mathematical axioms are necessarily false, we will need to think of the “worlds” in question as metaphysically impossible worlds. 29 I take it that the hypothesis that matter is all alike is (a priori) consistent with any con- sistent hypothesis about the distribution of matter in space and time. If the hypothesis that there is some matter shaped in a certain way cannot be ruled out a priori, then neither can the hypothesis that some homogeneous matter is shaped in that way. 30 Provided that the original theory T posits points and regions of fl at space-time, we can use an extension of this method to eliminate its commitment to subatomic particles alto- gether. To begin with, we must supplement T by adding “bridge laws” to make explicit the ways in which the facts about the positions and motions of atoms depend on the facts about subatomic particles. Next, we reconstrue all talk of subatomic particles as talk of regions of space-time: instead of saying that a particle is present at a certain space-time point, we say that the point is a part of the region that is the particle. Finally, as before, we replace predicates like ‘electron’ that purport to pick out different kinds of space-time regions with variables ranging over sets of space-time regions, and add initial existential quantifi ers to bind these variables. The resulting theory will have the same consequences as the original theory as regards the spatiotemporal distribution of atoms, but will be consistent with the hypothesis that the only concrete things there are, fundamentally speaking, are atoms moving around in a space-time manifold that is otherwise completely homogeneous. But this hypothesis entails that there aren’t any subatomic particles, even in a superfi cial sense. Perhaps subatomic particles could exist, superfi cially speaking, in a world where there were only space-time points and atoms, fundamentally speaking. But this could happen only in virtue of some asymmetries in the intrinsic character of the space-time, such as electromagnetic fi elds, or variable curvature. 31 Indeed, once we have enough mathematics on board, we can always use existential quanti- fi ers instead of possibility-operators like the one in T †. Instead of saying that the facts about the positions and motions of atoms are consistent with T, we can say that there is some model of T that accurately represents the facts about the positions and motions of atoms. 32 Gauge theories provide an even more dramatic example. See Belot (1998). 33 For the record, the view I favor is one on which the analyses of counterfactuals will typi- cally be statable only in a language that allows for infi nitely long sentences. 34 Cf. Fine (2001). 35 Armstrong also holds that the properties that feature in the analyses of these elite predi- cates are the only properties there are, fundamentally speaking. But even a believer in “abundant” properties could agree with Armstrong that only a few of these properties feature in true instances of (11). 36 I should note that Armstrong believes in “structural” properties – e.g. a property being a helium atom that is necessarily instantiated by any atom whose nucleus contains exactly two protons. And he seems to endorse (11) for the predicates corresponding to such prop- erties. Thus, unlike me, he apparently sees no diffi culty in accepting both (15) and (16). 37 Armstrong tends to focus instead on predicates like ‘exactly resembles in some respect’ (proposed analysis: ‘has some property in common with’), and ‘is a natural class’ (proposed analysis: ‘has as members all and only the things that instantiate some property’). But it is not clear that a nominalist should regard these predicates as precise enough to admit of any easily stated analysis. 38 If one rejects Armstrong-style “structural” properties (see note 36), and accepts the exis- tence in the fundamental sense of composite objects (contrary to the position defended in Dorr 2005), one will require a more complicated analysis of ‘duplicate’ than this. The problem with (18) is that composite objects can fail to be duplicates by having parts that There Are No Abstract Objects 57 fail to be duplicates, even though they themselves may not instantiate any real properties. Here is an alternative analysis that avoids this problem: for x and y to be duplicates is for there to be a one-to-one correspondence (bijection) between the parts of x and the parts of y, such that corresponding parts instantiate the same properties and stand in the same relations (cf. Lewis 1986: 61–2), 39 One might refuse to accept that (D) is necessary on the grounds that it might not even be true. If things get to be electrons in virtue of how they stand to other things – e.g. in virtue of being lower in charge than other particles (Lewis 1986: 76) – there is nothing to stop a non-electron from being a duplicate of an electron. The easiest way to respond to this worry is to weaken (D) by replacing the notion of duplication with that of qualita- tive indiscernibility (Lewis 1986: 63) – exact resemblance in extrinsic as well as intrinsic respects, of the sort that can occur only if the whole world is perfectly symmetric. 40 Devitt (1980), Field (1992), Van Cleve (1994), and Melia (2005) all defend views that might be interpreted as versions of the physical strategy. 41 As with the analysis of duplication in terms of instantiation (note 38), we will need a more complicated analysis to handle duplication among complex objects. Here is one way we might do it:
For x to be a duplicate of y is for there to be a bijection f from the parts of x to the parts of
y, such that for all parts z1, z2, z3 . . . of x: z1 is an electron iff f (z1) is an electron, . . . and z1 is
more massive than z2 iff f (z1) is more massive than f (z2), . . . and z1 is between z2 and z3 iff
f (z1) is between f (z2) and f (z3), and . . .
Here the quantifi ed formula has one conjunct for each basic physical predicate. Of course, this will not be acceptable as it stands to a nominalist, given the use of quantifi cation over functions. But we have already seen, in section 2, how such quantifi cation can be paraphrased in such a way as to be acceptable to the nominalist. 42 Ironically, Armstrong (1989a) once argued for the claim that alien properties are impossi- ble. But he later changed his mind about this (1997: section 10.4). 43 Notice that some of the modal intuitions that confl ict with an analysis like (19) can be stated without using physical predicates like ‘electron’ at all. For example, an analysis using exactly 17 basic physical predicates will entail, counterintuitively, that it is impos- sible for there to be 217 + 1 simple objects none of which is a duplicate of any of the others. This may be important. Several philosophers (e.g., Bealer 1987; Chalmers 1996) have attempted to draw a principled distinction between words like ‘water’ which can give rise to the phenomenon of the necessary a posteriori, and other words which cannot. But there is any such distinction to be drawn, it would seem that all the words in the sentence ‘it is not the case that there are 217 + 1 simple objects none of which is a duplicate of any of the others’ belong to the latter category. If so, the proposition expressed by this sentence could be necessary only if it were knowable a priori, as it clearly is not. 44 Or at least, anything qualitatively indiscernible from a white thing: see note 39. 45 For discussions of this view see Lewis (1983) and Armstrong (1989b). 46 For arguments, independent of nominalism, against the view that that plural claims of this kind can be analyzed in terms of sets, see Boolos (1984) and Oliver and Smiley (2001). 47 Although it is controversial whether anything that exists in the fundamental sense has any parts – see Dorr (2005). 48 For discussion of some of these complications, see Rodriguez-Pereyra (2002: 131–41). 49 Alternatively, we could allow all electrons to play the role of paradigms: this is the approach favored by Rodriguez-Pereyra (2002).
Cian Dorr 58 50 Armstrong (1978a: 50–1) makes essentially this objection. 51 One might attempt to escape this argument by saying something like (21*) To be an electron
is to resemble each of e1, e2 . . . as they actually are at least as much as they actually resemble one another. This is unproblematic for modal realists like Lewis (1986) and Rodriguez-Pereyra (2002). But any defense of nominalism that relied on modal realism would be widely, and in my view correctly, regarded as a reductio. For a non-modal-realist, “cross-world” similarity can hardly be primitive. One could attempt to analyze it in terms of objects playing similar structural roles at their respective worlds: such an analysis would face the same diffi culties as the purely general structural analyses I will be considering in the main text. Ideas along these lines have been considered under the heading of ‘coun- terpart theory for properties’: see Hazen (1996), Heller (1998), Black (2000), and Hawthorne (2001). 52 Defenders of dispositional essentialism include Shoemaker (1980, 1998), Swoyer (1982), and Ellis and Lierse (1994). Hawthorne (2001) is an excellent survey of the arguments. For opposition, see Sidelle (2002). 53 Some (e.g., Swoyer 1982) go so far as to hold that all nomologically necessary truths are metaphysically necessary: this requires either rejecting the Alien Properties Intuition, or adopting an unusually restrictive notion of nomological necessity. 54 The problems physical symmetries pose for dispositional essentialism are extensively dis- cussed by Hawthorne (2001). 55 In the unlikely event that the electrons are structurally indiscernible from the positrons – if, for example, there is eternal recurrence in both directions, with every second epoch having positrons substituted for electrons – there will be no way to make the necessary distinc- tions at the purely structural level. But perhaps it wouldn’t be so bad to reject the Alien Individuals Intuition in such bizarre circumstances. 56 There is a helpful analogy between the problem the predicates ‘electron’ and ‘positron’ pose for the Structural Nominalist and the problem ‘left’ and ‘right’ pose for those who think that facts about leftness and rightness are not among the basic geometric facts about the world. We are tempted by the intuition that the other geometric facts don’t determine the facts about left and right: for example, that a hand in otherwise empty space could be either right or left. Despite this, the received view is that we need not expand our conception of the basic geometric facts to account for leftness and rightness. ‘Left’ and ‘right’ can be analyzed either using paradigms, or by appealing to some initially contingent-seeming asymmetries in the universe, the usual example is the fact that human beings’ hearts are on the left side of their bodies. For further discussion, see the papers collected in Van Cleve and Frederick (1991). 57 ‘Phlogiston’ was introduced by eighteenth-century chemists as a name for a hypothetical substance emitted in burning and the calcination of metals. See Dorr (2005: section 16) for a more detailed discussion of the case of semantically defective predicates and its ramifi cations. 58 Blackburn (1993) and Gibbard (1990) are the most prominent evaluative non-factualists. For a defense of the distinction between the factual and the non-factual and some inter- esting ideas about how to draw it, see Fine (2001). 59 The appeal to economy of brute necessities is one of Armstrong’s characteristic argumenta- tive moves – see, e.g., Armstrong (1978a: 49–50). Lewis (1983) is also explicit about the role of this sort of economy. 60 See Ramsey (1925), Armstrong (1978b: ch. 13), and Mellor (1991). 61 Ramsey (1925) seems to be proposing something like this, but with a symmetric primitive predicate ‘x and y constitute a fact’ in place of the non-symmetric ‘proto-instantiates’.
There Are No Abstract Objects 59 62 Assuming that the set of sentences that intuitively strike us as necessary is logically con- sistent with any logically consistent hypothesis according to which nothing falls under any of our old predicates. 63 See Goodman (1951: 86–9). 64 Ramsey (1925: 29), following Wittgenstein (1921), embraces the skeptical response to this objection. Since ‘we know and can know nothing whatever about the forms of atomic propositions’, we can’t know whether the fundamental objects divide into two different kinds that behave in the ways that might justify calling them ‘universals’ and ‘individuals’. 65 Even if nominalism is true, there are still necessary truths having to do with abstract entities whose necessity needs explaining. The most important example is the nominalist thesis itself, if it is taken to be necessary. It is certainly not obvious how to formulate analyses of predicates like ‘number’ and ‘property’ that would allow sentences like ‘there are no numbers’ and ‘there are no properties’ to be reduced to logical truths. The problem here seems quite similar to that posed by sentences like ‘there are no unicorns’ and ‘there are no phlogiston’: it is best dealt with, in my view, by making special allowance for “semantically defective” predicates in formulating the ban on brute necessities. 66 This is of course the merest sketch of an argument; I hope to fi ll in more of its details in future work. The biggest task in doing so is explaining why the logical truths, narrowly conceived, should be a better place for explanations of necessity to stop than any larger set of truths. 67 The hardest bullet to bite for those who regard basic physical predicates as primitive and unanalyzable while rejecting brute necessities is the contingency of even the most neces- sary-seeming geometric axioms. See Le Poidevin (2004). 68 Thanks to Hartry Field, Jessica Moss, Kieran Setiya, Ted Sider, and to audiences at Pittsburgh, Texas, and MIT.
References
Alston, William. 1963. “Ontological Commitments,” Philosophical Studies 14: 1–8. Armstrong, David M. 1978a. Nominalism and Realism. Vol. 1 of Universals and Scientifi c Realism (Cambridge: Cambridge University Press). ——. 1978b. A Theory of Universals. Vol. 2 of Universals and Scientifi c Realism (Cambridge: Cambridge University Press). ——. 1989a. A Combinatorial Theory of Possibility (Cambridge: Cambridge University Press). ——. 1989b. Universals: An Opinionated Introduction (Boulder: Westview Press). ——. 1997. A World of States of Affairs (Cambridge: Cambridge University Press). Bealer, George. 1982. Quality and Concept (Oxford: Oxford University Press). ——. 1987. “The Philosophical Limits of Scientifi c Essentialism,” in James Tomberlin, ed., Philosophical Perspectives 1: Metaphysics (Atascadero: Ridgewood), pp. 289–365. ——. 1993. “Universals,” Journal of Philosophy 110: 5–32. Belot, Gordon. 1998. “Understanding Electromagnetism,” British Journal for the Philosophy of Science 49: 531–55. Benacerraf, Paul. 1973. “Mathematical Truth,” Journal of Philosophy 19: 661–79. Black, Robert. 2000. “Against Quidditism,” Australasian Journal of Philosophy 78: 87– 104. Blackburn, Simon. 1993. Essays in Quasi-Realism (Oxford: Oxford University Press). Boolos, George. 1984. “To Be is To Be the Value of a Variable (Or To Be Some Values of Some Variables),” Journal of Philosophy 81: 430–449. Cian Dorr 60 Bradley, F. H. 1897. Appearance and Reality, 2nd edn. (Oxford: Oxford University Press). Repr. in Writings on Logic and Metaphysics, ed. James W. Allard and Guy Stock (Oxford: Oxford University Press), 1994. Burgess, John P., and Rosen, Gideon. 1997. A Subject with No Object: Strategies for Nominalistic Interpretation of Mathematics (Oxford: Oxford University Press). ——. 2005. “Nominalism Reconsidered,” in Stewart Shapiro, ed., The Oxford Handbook of Philosophy of Mathematics and Logic (Oxford: Oxford University Press), pp. 515–35. Campbell, Keith. 1991. Abstract Particulars (Oxford: Basil Blackwell). Carnap, Rudolf. 1950. “Empiricism, Semantics and Ontology,” Revue International de Philoso- phie 4: 20–40. Repr. in Meaning and Necessity: A Study in Semantics and Modal Logic, 2nd edn. (Chicago: University of Chicago Press, 1956). Chalmers, David. 1996. The Conscious Mind (Oxford: Oxford University Press). Chihara, Charles. 1990. Constructibility and Mathematical Existence (Oxford: Oxford University Press). Devitt, Michael. 1980. “ ‘Ostrich Nominalism’ or ‘Mirage Realism’?” Pacifi c Philosophical Quar- terly 61: 433–49. Repr. in D. H Mellor and Alex Oliver, eds., Properties (Oxford: Oxford University Press, 1997), pp. 101–11. Dorr, Cian. 2005. “What We Disagree About When We Disagree About Ontology,” in Mark Kalderon, ed., Fictionalist Approaches to Metaphysics (Oxford: Oxford University Press). Ellis, Brian, and Lierse, Caroline. 1994. “Dispositional Essentialism.” Australasian Journal of Philosophy 72: 27–45. Field, Hartry. 1980. Science Without Numbers (Oxford: Blackwell). ——. 1985. “On Conservativeness and Incompleteness,” Journal of Philosophy 81: 239–260. Repr. in Realism, Mathematics and Modality (Oxford: Blackwell, 1989), pp. 125–46. ——. 1988. “Realism, Mathematics and Modality,” Philosophical Topics 19: 57–107. Repr. in Realism, Mathematics and Modality (Oxford: Blackwell, 1989), pp. 227–81. ——. 1992. “Physicalism,” in John Earman, ed., Inference, Explanation, and Other Frustrations: Essays in the Philosophy of Science (Berkeley: University of California Press), pp. 271–91. Fine, Kit. 1994. “Essence and Modality,” in James Tomberlin, ed., Philosophical Perspectives 8: Logic and Language (Atascadero: Ridgeview), pp. 1–16. ——. 2001. “The Question of Realism.” Philosopher”s Imprint 1. Gibbard, Allan. 1990. Wise Choices, Apt Feelings. Cambridge, MA: Harvard University Press. Goodman, Nelson. 1951. The Structure of Appearance (3rd edn., Cambridge, MA: Harvard University Press, 1977). Hawthorne, John. 2001. “Causal Structuralism,” in James Tomberlin, ed., Philosophical Perspec- tives 15: Metaphysics (Cambridge, MA: Blackwell), pp. 361–78. Hazen, Allen. 1996. “Actualism Again.” Philosophical Studies 84: 155–181. Heller, Mark. 1998. “Property Counterparts in Ersatz Worlds.” Journal of Philosophy 95: 293–316. Hellman, Geoffrey. 1989. Mathematics Without Numbers (Oxford: Oxford University Press). Jackson, Frank. 1977. “Statements About Universals.” Mind 86: 427–429. Repr. in D. H Mellor and Alex Oliver, eds., Properties (Oxford: Oxford University Press, 1997), pp. 89–92. Kripke, Saul. 1972. Naming and Necessity. Rev. edn., Cambridge, MA: Harvard University Press, 1980. Le Poidevin, Robin. 2004. “Space, Supervenience and Substantivalism,” Analysis 64: 191–8. Lewis, David. 1973. Counterfactuals (Oxford: Blackwell). ——. 1983. “New Work for a Theory of Universals,” Australasian Journal of Philosophy 61: 343–77. Repr. in Papers in Metaphysics and Epistemology (Cambridge: Cambridge University Press, 1999), pp. 8–55. ——. 1986. On the Plurality of Worlds (Oxford: Blackwell). There Are No Abstract Objects 61 ——. 1993. “Mathematics is Megethology,” Philosophia Mathematica 1: 3–23. Repr. in Papers in Philosophical Logic. (Cambridge: Cambridge University Press: 1998), pp. 203–29. Melia, Joseph. 2005. “Truthmaking Without Truthmakers,” in Helen Beebee and Julian Dodd, eds., Truthmakers: The Contemporary Debate (Oxford: Oxford University Press). Mellor, D. H. 1991. “Properties and Predicates,” in Matters of Metaphysics (Cambridge: Cam- bridge University Press), pp. 170–82. Nolan, Daniel. 1997. “Impossible Worlds: A Modest Approach,” Notre Dame Journal of Formal Logic 38: 535–72. Oliver, Alex, and Smiley, Timothy. 2001. “Strategies for a Logic of Plurals,” Philosophical Quarterly 51: 289–306. Pap, Arthur. 1959–60. “Nominalism, Empiricism and Universals,” Philosophical Quarterly 9–10: 330–40, 44–60. Price, H. H. 1953. Thinking and Experience. 2nd edn., London: Hutchinson, 1969. Putnam, Hilary. 1967. “Mathematics Without Foundations,” Journal of Philosophy 64: 5–22. Repr. in Mathematics, Matter and Method: Philosophical Papers, vol. 1. (Cambridge: Cam- bridge University Press), pp. 43–59. ——. 1971. Philosophy of Logic. New York: Harper and Row. Repr. in Mathematics, Matter and Method: Philosophical Papers, vol. 1. Cambridge: Cambridge University Press, 1975, pp. 323–57. Quine, Willard van Orman. 1948. “On What There Is,” Review of Metaphysics 2: 21–38. Repr. with revisions in From a Logical Point of View (Cambridge, MA: Harvard University Press, 1953), pp. 1–19. ——. 1951. “Two Dogmas of Empiricism.” Philosophical Review 60: 20–43. Repr. with revisions in From a Logical Point of View (Cambridge, MA: Harvard University Press, 1953), pp. 20–46. Ramsey, F. P. 1925. “Universals,” Mind 34: 401–417. Repr. in D. H. Mellor, ed., Philosophical Papers (Cambridge: Cambridge University Press, 1990), pp. 8–33. Rodriguez-Pereyra, Gonzalo. 2002. Resemblance Nominalism (Oxford: Oxford University Press). Rosen, Gideon. 1990. “Modal Fictionalism,” Mind 99: 327–54. Russell, Bertrand. 1912. The Principles of Philosophy (London: Williams and Norgate). Schiffer, Stephen. 2003. The Things We Mean (Oxford: Oxford University Press). Shapiro, Stewart. 1983. “Conservativeness and Incompleteness,” Journal of Philosophy 80: 521–31. Shoemaker, Sydney. 1980. “Causality and Properties,” in Peter van Inwagen, ed., Time and Cause (Dordrecht: D. Reidel), pp. 228–54. ——. 1998. “Causal and Metaphysical Necessity.” Pacifi c Philosophical Quarterly 79: 59– 77. Sidelle, Alan. 2002. “On the Metaphysical Contingency of the Laws of Nature,” in Tamar Gendler and John Hawthorne, eds., Conceivability and Possibility (Oxford: Oxford University Press), pp. 309–36. Stalnaker, Robert. 1968. “A Theory of Conditionals,” in Studies in Logical Theory: American Philosophical Quarterly Monograph Series, No. 2 (Oxford: Blackwell). Stanley, Jason. 2001. “Hermeneutic Fictionalism,” in Peter French and Howard K. Wettstein, eds., Midwest Studies in Philosophy XXV: Figurative Language (Oxford: Blackwell), pp. 36–71. Swoyer, Chris. 1982. “The Nature of Natural Laws,” Australasian Journal of Philosophy 60: 203–23. Van Cleve, James. 1994. “Predication Without Universals? A Fling with Ostrich Nominalism,” Philosophy and Phenomenological Research 54: 577–90. Cian Dorr 62 Van Cleve, James, and Frederick, Robert E., eds. 1991. The Philosophy of Right and Left: Incon- gruent Counterparts and the Nature of Space (Dordrecht: Kluwer Academic Publishers). van Fraassen, Bas. 1980. The Scientifi c Image (Oxford: Oxford University Press). van Inwagen, Peter. 1977. “Creatures of Fiction,” American Philosophical Quarterly 14: 299– 308. Repr. in Ontology, Identity and Modality (Cambridge: Cambridge University Press, 2001), pp. 37–56. ——. 2004. “A Theory of Properties,” in Dean Zimmerman, ed., Oxford Studies in Metaphysics, vol. 1 (Oxford: Oxford University Press), pp. 107–38. Wittgenstein, Ludwig. 1921. Tractatus Logico-Philosophicus. English edn., ed. D. F. Pears and B. F. McGuinness. (London: Routledge and Kegan Paul, 1961). Wright, Crispin. 1983. Frege’s Conception of Numbers as Objects (Aberdeen: Aberdeen University Press). Yablo, Stephen. 2000. “A Paradox of Existence,” in Anthony Everett and Thomas Hofweber, eds., Empty Names, Fiction and the Puzzles of Non-Existence (Stanford: CSLI Press), pp. 275–312. ——. 2001. “Go Figure: A Path Through Fictionalism,” in Peter French and Howard K. Wettstein, eds., Midwest Studies in Philosophy XXV: Figurative Language (Oxford: Blackwell), pp. 72–102.
There Are No Abstract Objects 63
CHAPTER TWO CAUSATION AND LAWS OF NATURE
2.1 “Nailed to Hume’s Cross?” John W. Carroll 2.2 “Causation and Laws of Nature: Reductionism,” Jonathan Schaffer
It just happened to be true, let us suppose, that everyone who ate at the Mar-T Café on April 8, 1990, wore a blue shirt. Other events are not so “accidental.” For example, it’s no accident that when the cook let go of the French fries, they fell into the fryer. In some sense, the fries had to fall, given that the cook let them go. When an event is caused, and when there is a law of nature governing its occurrence, it is in some sense necessary that the event occurs. Where does this necessity come from? Jonathan Schaffer argues that the necessity boils down to mere regularities. The necessity of the fries’ falling boils down to the fact that fries everywhere, and every time, in fact do fall when they are released. John Carroll argues that there is more to it than this; causal and lawful necessity go beyond mere regularity.
CHAPTER 2.1
Nailed to Hume’s Cross?
John W. Carroll
1 Lawhood, Causation, and Bearing Hume’s Cross
Some scientists try to discover and report laws of nature. And, they do so with success. There are many principles that were for a long time thought to be laws that turned out to be useful approximations, like Newton’s gravitational principle. There are others that were thought to be laws and still are considered laws, like Einstein’s principle that no signals travel faster than light. Laws of nature are not just important to sci- entists. They are also of great interest to us philosophers, though primarily in an ancillary way. Qua philosophers, we do not try to discover what the laws are. We care about what it is to be a law, about lawhood, the essential difference between something’s being a law and something’s not being a law. It is one of our jobs to understand lawhood and convey our understanding to others. Causation is also central to science and to philosophy. Molecular bonding, plane- tary orbits, human decisions, and life itself are all causal processes. A scientifi c explanation of an event will include some mention of the causes of that event – you can’t say why something did happen without identifying what made it happen. Just as is the case for lawhood, qua philosophers, one job we have is to understand causa- tion and then to share this understanding with others. As a result of the work of David Hume, many philosophers are infl uenced by a metaphysical concern and a skeptical challenge that have shaped what is counted as providing understanding of lawhood and causation. Hume’s argument against the idea of necessary connection contains the plausible premise that we lack any direct per- ceptual or introspective access to the causal relation:
All events seem entirely loose and separate. One event follows another, but we never can observe any tie between them. They seem conjoined, but never connected. But as we can have no idea of anything which never appeared to our outward sense or inward sentiment, the necessary conclusion seems to be that we have no idea of connection or power at all, and that these words are absolutely without any meaning when employed either in philosophical reasonings or common life. (1955: 85; italics in the original)
The skeptical challenge that emerges says that, if lawhood and causation are not analyzable in terms of some more accessible notions, then we would be prevented from having knowledge we ordinarily take ourselves to have. Regarding the meta- physical concern, for various reasons, in trying to say what makes it true that some- thing is a law or that one thing causes another, it can be tempting not to limit oneself to the accessible notions, instead positing necessary connections or other questionable entities as existing in the world. Short of doing so, the concern is that the truthmakers for reports of lawhood and causation would be non-existent, and then the reports couldn’t be true; for Hume, they couldn’t even have meaning. To address these worries, philosophers often seek a certain sort of analysis of lawhood. They seek a necessarily true completion of:
(S1) P is a law of nature if and only if . . .
The expectation is that the analysis should make clear that lawhood is suitably acces- sible for us to have the knowledge of laws we ordinarily take ourselves to have and make clear what it is about the universe that makes true reports of lawhood true. There has been widespread agreement that certain sorts of completions of (S1) are unsatisfactory in this regard. Analyses of lawhood that use the counterfactual condi- tional (i.e., if P were the case, then Q would be the case) would do little to address the Hume-inspired worries. Concerns about our knowledge of counterfactuals and their truthmakers are just as prevalent as are the parallel worries about lawhood. The same can be said for completions of (S1) that employ the other nomic concepts: cau- sation, lawhood, explanation, chance, dispositions, and their conceptual kin. As a result, the history of philosophy has shown a preference for what I call a reduction of lawhood; philosophers have tried to provide a necessarily true completion of (S1) without using any nomic terms. The history of philosophy includes many attempts to give a similarly reductive analysis of causation, though, at least recently, the constraints on what counts as a satisfactory analysis of causation have been somewhat less severe. There is still much preoccupation with giving an analysis of causation, with fi nding some necessarily true completion of:
(S2) P caused Q if and only if . . .
And no, not just any necessarily true completion will do; for example, no one would bother with P caused Q if and only if P caused Q. But, unlike with lawhood, there has been a lot of attention given to analyzing causation in terms of chance, the counterfactual conditional, lawhood, or some combination of these nomic concepts. That is, in the last 40 years or so, philosophers have not insisted that (S2) be completed non-nomically. This difference in attitude is easy to explain. An underlying belief of the philosophical community has been that lawhood is the best place to get off the nomic bus and squelch the Hume-inspired worries. The thought seems to be that, as John W. Carroll 68 a practical matter, it is easier to give a thorough reduction of lawhood. So, we are better off analyzing causation, say, in terms of the counterfactual conditional, maybe analyzing this conditional in terms of lawhood, and then letting a non-nomic analysis of lawhood do the last bit of reductive work. In Laws of Nature (1994), I argue that the history of philosophy has been pretty badly wrong about all of this. I maintain that neither causation nor lawhood can be analyzed non-nomically and, further, that causation even resists any (non-circular) analysis in terms of the counterfactual conditional, chance, or lawhood. What I propose to do in this chapter is defend my brand of anti-reductionism against the Hume-inspired worries. So, after wrapping up this introductory section, I will quickly review in section 2 an example that challenges the prospects of giving a Humean reduction of lawhood, a reduction of lawhood that does not require that we posit any mysterious ontology. (This example will at the same time also provide a nice basis for presenting certain other ways of trying to understand what it is to be a law.) In section 3, I take on the metaphysical concern as it applies to my anti-reductionism, offering a sketch of a new non-reductive theory of lawhood. In section 4, I use this analysis to shed light on the skeptical challenge. In one regrettable way, my defense of anti-reductionism will be limited. I will only directly address the Hume-inspired concerns as they apply to lawhood – defending my anti-reductionism about causation will have to wait. Given the diffi culty of the two tasks, I cannot do a good job on both, and it seems to me to be more appropriate to focus on lawhood rather than causation because more attention has been paid in recent years to giving a thoroughly reductive account of lawhood. This is not to say that this chapter is not about causation. As we shall see, my non-reductive theory is non-reductive precisely because it invokes causation (or, strictly speaking, a closely related notion of explanation). I will argue that the laws of nature are exactly those regularities that are caused by nature.
2 Support for Anti-Reductionism and a Glance at Some Alternatives
Suppose that there are exactly 10 different kinds of fundamental particle. So there are 55 possible kinds of two-particle interaction. Suppose also that 54 of these kinds of interaction have been studied and 54 laws have been proposed and thoroughly tested. It just so happens that there are never any interactions between the last two kinds of particle; these are arbitrarily labeled as ‘X’ and ‘Y’ particles. One thing that is ingenious about this example of Michael Tooley’s (1977: 669) is that, at least at fi rst glance, it seems that there could be a law about X–Y interactions. After all, in the example, scientists have already discovered laws for all of the other 54 kinds of interaction. Indeed, it is even true that some of these laws are about X particles and that some are about Y particles. Thus, there seems to be some reason to think that there is also a law about what will happen if X and Y particles get together. Another thing that is ingenious about Tooley’s case is that it seems that many different X–Y interaction laws are perfectly consistent with all the events that might take place during the complete (i.e., past, present, and future) history of the 10-particle world. It seems Nailed to Hume’s Cross? 69 that the totality of events of this universe could fail to determine what the laws are.
Even given the complete history of this universe, there might be a law, L1, that, when X particles and Y particles interact, the particles are destroyed. But, then again, even
given the complete history of this universe, there might be a law, L2, that, when X par- ticles and Y particles interact, the particles bond. It seems that what the laws are does not supervene on (i.e., is not determined by) the non-nomic facts. Tooley takes his example to make a case against Humean reductive attempts to solve the problem of laws. To see why, consider what such an account might say about his 10-particle world. For example, a naive Humean might hold that P is a law of nature if and only if P is a true, contingent, universal generalization. This account says about Tooley’s example that both L1 and L2 are laws: it is a law of the 10-particle world that any interaction of X and Y particles results in their annihilation and it is a law of that world that when X and Y particles interact they bond. But that is impos- sible because such annihilation and bonding events are incompatible; it cannot be true that, if there were an X–Y interaction, then there would be both the bonding and the annihilation. This problem for the simplistic Humean account is a consequence of the fact that the account does not differentiate between L1 and L2. Because the two are both true, contingent, generalizations, they both get counted as laws. David Lewis (1973, 1983, and 1986) holds a much more sophisticated Humean account, one that maintains that P is a law of nature if and only if P is a member of all the true deduc- tive systems with a best combination of simplicity and strength. But, at least at fi rst glance, his view is faced with the same problem that faces the naive Humean view.
Not only do L1 and L2 not differ regarding their logical form, their contingency, or their truth, they also do not differ regarding their simplicity or their strength. Prima facie, either they both would belong to all the best systems or neither would. Humean reductionists must somehow deny that the 10-particle case is genuinely possible. In contrast, a universalist reductive approach of the sort favored by Tooley and also by David Armstrong (1983) and Fred Dretske (1977) seems to be in better shape. Much simplifi ed, universalists hold that: Fs are Gs is a law of nature if and only if the universal F-ness stands in N to the universal G-ness. (‘N’ and sometimes ‘necessitation’ are names given to the two-placed relation that relates universals. When F-ness stands in N to G-ness, F-ness is said to necessitate G-ness.) In virtue of its appeal to facts about universals, this is not a Humean reduc- tion of lawhood. Just so, it leaves open the possibility that there is something that grounds the lawhood of exactly one of the generalizations in Tooley’s 10-particle case. For all that has been said, it might be the case that being an X–Y interaction necessitates annihilation but not bonding. It could also go the other way: being an X–Y interaction might necessitate bonding but not annihilation. We need not draw the conclusions drawn by the universalists. Armstrong, Dretske, and Tooley are clearly still very much stuck with the Hume-inspired skeptical chal- lenge. Prima facie, identifying the truthmaker for lawhood reports with a relation (itself taken to be a universal) that holds between universals does nothing to make it clear how observational data could support knowledge of what the laws are. Further- more, despite appearing to have identifi ed a reductive truthmaker for lawhood reports, John W. Carroll 70 the metaphysical concern really is still an issue for the universalists. Tooley’s example exposes a void in the universalist approach, at least insofar as that view has been presented here. It is awfully convenient that the universals line up so nicely in Plato’s Heaven, doling out lawfulness to exactly the regularities that are laws. But, why should we believe that they really do line up so nicely? What do we really know about this necessitation relation, besides its name? What relation is it? Without a specifi cation of the relation, the universalists have not really given a reductive analy- sis of lawhood. They haven’t given an analysis period. Some philosophers react to the 10-particle case differently from the Humean reduc- tionists and the universalists. At a loss to identify the missing truthmakers, or troubled by the skeptical challenge, they conclude that lawhood sentences do not describe reality. They are anti-realists about lawhood. Some anti-realists, e.g., Bas van Fraassen (1989) and Ronald Giere (1999), go so far as to assert that there are no laws. Other phi- losophers, e.g., Simon Blackburn (1984, 1986) and Barry Ward (2002, 2003), adopt a different sort of anti-realism. Though they will utter sentences like, ‘It is a law that no signals travel faster than light’, they are anti-realists in virtue of thinking that the purpose of such sentences is not to describe the facts. On their view, lawhood sentences convey no information over and above what is conveyed by their contained generaliza- tion sentence. Instead, these sentences project a certain non-cognitive attitude about the conveyed generalization. All anti-realists sidestep the Hume-inspired skeptical and metaphysical worries: if lawhood reports are not even meant to describe the way our world is, or do have that purpose but are all false, then we certainly do not need to worry about their truthmakers or how we could know what the laws are. Certain features of my own view stand out when contrasted with the other standard positions on lawhood. I am an anti-reductionist in denying that there are any neces- sarily true Humean completions of (S1), but also in denying that there is any onto- logically rich non-Humean reduction of the sort defended by the universalists. I am also an anti-supervenience theorist because I accept Tooley’s 10-particle case at face value, and indeed use examples like it in my book Laws of Nature as the centerpiece of my arguments for anti-reductionism. So, as I see it, the non-nomic does not deter- mine what the laws are: there are possible worlds that agree on their non-nomic facts and disagree about what the laws are. Yet I am also a realist; I do think that there are utterances of lawhood sentences that try (and even sometimes succeed) in describ- ing reality – it is true and I know it is true that it is a law that no signals travel faster than light. Thus, my realist, anti-supervenist, anti-reductionism seems to put me in a horrible position in relation to the Hume-inspired worries. On my view, it is true that there are laws and some of us know what some of the laws are, but there is no reductive analysis of lawhood that could explain how this knowledge of lawhood is attainable or even say what makes it true that the laws are the laws. What can a realist, anti-supervenist, anti-reductionist say to Hume?
3 The Metaphysical Concern
Sometimes the metaphysical concern takes the form of a worry about how uninforma- tive anti-reductionism about lawhood is bound to be. Tooley has this to say about an Nailed to Hume’s Cross? 71 account of lawhood offered using dispositional terms: “[I]n offering this sort of answer one is not really making any progress with respect to the problem of explaining nomological language in the broad sense” (1987: 68). This lack of progress, however, does nothing to establish even a presumption in favor of reductionism. True, the anti-reductionist denies that there is a reductive answer to the question of what makes something a law. But, failure to provide that sort of answer cannot count against the anti-reductionist and in favor of the reductionist without begging the question. As the anti-reductionist sees it, it is the reductionist that commits the transgression in giving a reductive answer to a question that does not have one. In any case, questions about informativeness aren’t really what scare philosophers away from a realist anti-reductionism. What does that is a demand for truthmakers. When saying why he believes that there needs to be a relation of necessitation between F-ness and G-ness for Fs are Gs to be a law, Armstrong makes such a demand:
Suppose that one is a Nominalist in the classical sense of the term, one who holds that everything there is is a particular, and a particular only. Because there is nothing identi- cal in the different instantiations of the law, such a Nominalist, it seems, is forced to hold a Regularity theory of law. For if he attempts to hold any sort of necessitation theory, then he can point to no ontological ground for the necessity. He is nailed to Hume’s cross. (1983: 78)
If we are unable to say what makes a generalization a law, then, especially given Tooley’s 10-particle example, it can appear that nothing does. What could it be that makes it true that L1 is a law rather than L2? What makes it true that L2 is a law rather than L1? What the laws are fl oats on nothing (cf., Armstrong 1983: 31). It is important to recognize that anti-reductionists about lawhood need not be primitivists. The standard arguments for anti-reductionism leave open that there might be some analysis of lawhood in terms of some other nomic concepts. For example, Marc Lange (2000) has analyzed lawhood in terms of membership in a counterfactu- ally stable set of propositions. Lange’s idea is that the set of laws is special in that each of its members would still be true under any counterfactual supposition consis- tent with the set itself. Prima facie, this analysis makes substantive and interesting claims about what lawhood is. We can make enough independent judgments about which counterfactuals are true and about which propositions are laws to test his pro- posal. If the analysis is correct, it speaks to any leftover concerns stemming from the supposed uninformativeness of anti-reductionism. Obviously, though, Lange’s approach will do nothing to convince anyone to set aside Armstrong’s fl oat-on-nothing worry. Counterfactuals live too close in logical space to lawhood and behave in all too similar ways. About Tooley’s example, what would make it true that, if some X particle were to interact with some Y particle, then an annihilation event would occur? What would make it true that, if some X particle were to interact with some Y particle, then a bonding event would occur? Nothing, it seems, would make either of these counter- factuals true. I am not endorsing the truthmaker concern. Anyone who has a primitive concept playing some role in their metaphysics (i.e., everyone with a metaphysics) has some- thing that in a certain sense lacks a truthmaker of the sort Armstrong demands
John W. Carroll 72 regarding lawhood. And, it has always seemed to me, and still does, that the coun- terfactual conditional is a pretty good candidate to be primitive in a respectable metaphysics. So, I am not moved to abandon my anti-reductionism, and Lange will not be moved to abandon his either. Nevertheless, I will propose an alternative analysis of lawhood, one open to anti-reductionists, that may speak to the metaphysical concern in a way that Lange’s non-reductive analysis does not. It is often alleged, though primarily in informal discussions of our topic, that no laws are accidentally true. Such remarks stem from two different sources: their plau- sibility and the feeling among philosophers that such remarks are no big deal because this is not where any real work will get done. Prima facie, the metaphysical and skeptical worries about lawhood apply in a straightforward way to non-accidentality, and this notion is no better understood than lawhood itself. Investigating accidentality is usually not considered a step in the right direction. Well, I say that, in this case, we have underestimated the power of an intuitive gloss. Idea: What is an accident equates with what is a coincidence, where a coinci- dence is something that is unexplained.1 I run into an old friend at a Durham Bulls game. I did not even know he was in North Carolina; he moved away more than three years ago. Our meeting at the game is a coincidence. What makes it a coincidence? Well, it is a coincidence because it just happened. In other words:
P is a coincidence if and only if there is no Q such that P because Q.
The key notion here is the one expressed by ordinary uses of ‘because’. Strictly speaking, it is a kind of explanation. It is, however, different from causation in only uninteresting ways. If b’s being F caused c’s being G, then c was G because b was F. The other direction is not as straightforward. The number 3 is a square root of 9 because 32 is 9, but we are reluctant to say that 32’s being 9 caused 3 to be a square root of 9. In general, we are reluctant to take mathematical explanations, explanations underwritten by defi nitions, and explanations involving universal generalizations to be causal ones. For some reason, when the explanandum and the explanans are too closely connected, connected in some more-or-less analytic way, and not by some paradigmatically causal process (e.g., colliding), or when the explanandum or explan- ans themselves are suffi ciently unlike paradigmatic causes and effects (e.g., moving billiard balls), philosophers tend not to consider these explanations to be causal. That, however, seems to me to be the extent of the difference. Causation and the relevant notion of explanation amount to pretty much the same thing. So, unoffi cially, and a bit more stylishly, I like to say that P is a coincidence if and only if P is uncaused. One might object that, despite what I said, my seeing my friend at the ball game has an explanation, lots of them even, and that it certainly wasn’t uncaused. Wasn’t the meeting the result of each of us deciding to take in a ball game over the weekend? Didn’t we run into each other at that game because we both like baseball? Fair enough, but also notice how strange it is to pair these explanations with the attribution of coincidentality. For example, notice how odd and even contradictory it is to say all in one breath: we met because we both like baseball and that we met was a coinci- dence. Insofar as our both liking baseball does explain why my friend and I met, the
Nailed to Hume’s Cross? 73 meeting was not a coincidence. As I see it, ‘because’ utterances are context-sensitive and their context sensitivity carries over to ‘coincidence’ utterances. In an ordinary context, like the one present when I fi rst introduced the baseball example, the sentence ‘My friend and I met at the game because we both like baseball’ is false. It is a bit of a long story (see my 2005), but such utterances are false because, without a context change, that we both like baseball is not suffi cient for my friend and me to meet at the game. In such a context, that we both like baseball, together with what was presupposed or common ground in the context, does not entail that we would meet. Context can shift so that such utterances are true, but then an utterance of ‘Our meeting at the ball game was a coincidence’ would not be true in the new context. What about laws? Laws are not coincidences. They are not things that just happen; they are explained.2 Not being a coincidence, however, is not all there is to being a law. For example, some particular states of affairs, like there being tobacco in North Carolina, are not coincidences but are not suitably general to be laws. Or, for a more interesting case, it might be true that there are no gold spheres greater than a mile in diameter because there is not enough gold. In that case, strictly speaking, it would be true, suitably general and not a coincidence that all gold spheres are less than a mile in diameter. Nevertheless, that still would not be a law; it is not enough to be a law to be general and not coincidental. What seems important about this gold- spheres example is how the regularity turns out not to be non-coincidental. What blocks it from being a law is that something in nature, or really a certain sort of initial condition of the universe, an absence of something in nature, explains the regularity. Contrast this with the law that no signals travel faster than light. With this generalization it seems that it is true because of nature itself. Lawhood requires that nature itself – understood as distinct from anything in nature or the absence of something from nature – make the regularity true.
P is a law of nature if and only if P is a regularity caused by nature.
While this is a catchy way to put my favored analysis of lawhood, there are certain aspects of my view that require comment. First, we should keep in mind the point made earlier about explanation and causation. My offi cial view is not that laws are caused by nature but that they hold because of nature. Second, and more important, self-respecting metaphysicians will surely ask what exactly nature is. Think of nature as the universe – not the objects and events in the universe, but whatever it is that the objects and events are in. Along this same line, we can think of nature as some- thing like the universe’s space-time manifold or the totality of its space and time. Better yet, think of nature as something like an omnipresent and eternal fi eld, a big- as-big-can-be magnetic fi eld that is also as longlasting as longlasting can be whose effects need not have anything to do with magnetism. On my view, a scientist who posits that there are laws of nature is thereby committed to our world being causal/ explanatory in exactly this way. Some will object to the idea that something like nature can stand in the causal/ explanatory relation that is being employed in my analysis. Nature is not an event. It is also not a state of affairs (i.e., an object or event having some property). Yet, it John W. Carroll 74 is a long-standing opinion of many metaphysicians that only events or states of affairs can cause anything. To some it may even sound like I am taking seriously the idea of substance causation, an idea that is often in disrepute.3 Nature is not a substance, exactly. It is more like a humongous and ancient fi eld – it contains objects and other substances, but is not itself one. Admittedly, nature is more like a substance than it is like an event or a state of affairs and that will still worry some. But taking it to be causal really should not be any more worrisome than thinking of a magnetic fi eld as causing an electron to move. Furthermore, my account leaves room for properties to play a role. Nature does cause what it does in virtue of being a certain way. On my view, the job of scientists set on discovering what regularities are laws of nature is precisely to describe what these properties are. Roughly, stating that P is a law is science’s way of describing how nature is in virtue of which it causes P. I do not here provide anything like a full defense of my analysis of lawhood. It will suffi ce for my purposes if I have said enough to make it seem plausible. My analysis is not reductive; a notion of causation/explanation expressed by the word ‘because’ occurs in the analysans, and that concept is a nomic one. Nevertheless, there is no circularity and the analysis provides understanding. Besides being informative, my analysis also seems to metaphysically ground lawhood in ways that non-reductive analyses normally do not. Of all the nomic concepts, causation seems the most grounded, the one that seems the least to fl oat on nothing. It does not seem to fl oat at all. It is a relation that only holds between existing, occurrent, or obtaining things. Sometimes we can scientifi cally describe an underlying causal mechanism when the causal relation is instantiated. Some authors have even argued that, despite Hume, causation is directly observable.4 It certainly seems right that I can see that Nomar hit the ball and it is clear that this fact is causal; Nomar couldn’t have hit the ball unless the ball moved because of Nomar. I am less sympathetic to there being some causal percept or an impression of causation in Hume’s sense, but I do not see that this matters. We make causal and explanatory judgments easily and without much thought all the time. And, yet, almost all being a law of nature amounts to is holding because of nature.
4 The Skeptical Challenge
To my mind, the most careful and the most confounding formulation of the skeptical challenge comes from John Earman and John Roberts, in their paper “Contact with the Nomic: A Challenge for Deniers of Humean Supervenience about Laws of Nature” (2005). Their paper will be the focus of my reply.
4.1 The challenge from Earman and Roberts Let T be a theory that posits at least one law. Label one of the laws ‘L’ and reformulate T as the conjunction that L is a law of nature and X. (So, X is the rest of T aside from the part of T that posits L as a law.) Let T* be the theory that L is true, L is not a law, and X. T and T* cannot both be true because they differ on whether L is a law. Nailed to Hume’s Cross? 75 The argument is straightforward:
(1) If HS (Humean Supervenience) is false, then no empirical evidence can favor T or T* over the other. (2) If no empirical evidence can favor T or T* over the other, then we cannot be epistemically justifi ed in believing on empirical grounds that T is true. ______(3) If HS is false, we cannot be epistemically justifi ed in believing on empirical grounds that T is true.5
HS is defi ned by Earman and Roberts as the thesis that what is a law and what is not cannot vary between worlds with the same Humean base, where a world’s Humean base is the set of non-nomic facts at that world that are detectable by a reliable measurement or observation procedure (p. 253). It follows from the premises that, if HS is false, then we cannot be justifi ed in believing in T. That is an apparent problem for my anti-supervenience, realist, anti-reductionism because Earman and Roberts made no assumptions about T other than that it attributes lawhood to at least one proposition. So, if Earman and Roberts’ argument is sound and HS is false, then no one is justifi ed in believing on empirical grounds that any proposition is a law. It is only a short step from there to the conclusion that no one – not us, not the scientists – know what any of the laws are. As Earman and Roberts are averse to skepticism, they ultimately see this as an argument for HS.
4.2 Empirical evidence against cosmic coincidences To begin my response, I will describe one basic way empirical evidence can support the judgment that something is a law. It is a way suggested by the non-reductive analysis of lawhood given in section 3. But, be warned. Contexts for utterances using the verbs ‘to know’ and ‘to justify’ are fragile. Without a lot of work, Earman and Roberts could (and may have already) spoiled the present context in such a way that some of the epistemological claims I am about to make will not ring true. That is why it will be important, in the next subsection, to say something about the context dependence of epistemological terms. Here is John Foster’s insightful description of a hypothetical case of an inference to there being a law. It is in line with the picture I want to sketch:
The past consistency of gravitational behavior calls for some explanation. For given the infi nite variety of ways in which bodies might have behaved non-gravitationally and, more importantly, the innumerable occasions on which some form of non-gravitational behavior might have occurred and been detected, the consistency would be an astonish- ing coincidence if it were merely accidental – so astonishing as to make the accident-hypothesis quite literally incredible. (1983: 89)
In this spirit, regarding Earman and Roberts’ argument, I want to suggest that believ- ing T* sometimes would be to believe implausibly that L does not hold because of nature. Since Earman and Roberts grant that we may have reason to believe L is true, John W. Carroll 76 believing T* means believing either that L is caused by nothing (and so is a coinci- dence) or else that L holds because of something in nature. Sometimes we have good empirical evidence against each of these disjuncts, evidence that thus favors T over T*, and so we may be justifi ed in believing T. It is not mysterious how we justifi edly judge whether certain generalizations are coincidences. This is especially true about some very local regularities. Suppose I come upon a bag of marbles. I open it up, peek in, and see that all the marbles in the bag (right now) are black. I am not likely to take seriously in this situation the possibility that nature explains why the generalization holds. Indeed, I am likely to presuppose that this is not the case. Still, it might be that there is some explanation of that gen- eralization. For instance, it might be that the marbles were selected from an urn that contained only black marbles. In contrast, it might also be that nothing explains why all the marbles in the bag are black. It might be that the marbles were selected blindly from an urn that contained a mix of black and white marbles, a mix with many more white marbles than black marbles. The important point to notice about these two kinds of possibilities – one kind that includes an explanation for the generalization and the other kind that does not – is that empirical evidence can favor one kind over the other. I might have seen someone picking the balls from an urn containing only black marbles. Then again, I might have seen someone blindly selecting the marbles from an urn with only a small proportion of black marbles. So, on the local scale anyway, there are straightforward ways of gaining evidence that would decide whether a regularity was a coincidence or instead was explained. With cosmic regularities, we are more likely to take seriously the idea that they might be caused by nature. Indeed, I suspect that physicists dealing with fundamental particles and properties are likely to presuppose that, if a generalization of interest to them is true, then it is not a coincidence, and so must be the result of something in nature or nature itself. Consider the principle of the conservation of energy. Years of investigation and careful theoriz- ing reveal that it has no in-nature explanation. The absence of any in-nature explanation supports the hypothesis that this principle holds because of nature, and so is a law. What is important is that sometimes we fi nd no in-nature explanations of a regu- larity, but we are also reluctant to conclude that it is a coincidence. We are faced with the choice of its being nature that explains it or its being unexplained. Sometimes the latter fi ts better with the rest of what we believe. When it does, we are justifi ed in believing a proposition corresponding to Earman and Roberts’ T *. With the right sort of empirical evidence, however, the coincidence hypothesis may be much less credible than the lawful one. So, we may be justifi ed in believing a proposition exactly parallel to Earman and Roberts’ T. In short, we sometimes have evidence of what nature causes and that is all the evidence we need to distinguish laws from non-laws.
4.3 Contextualism and relevant alternatives My guess is Earman and Roberts will disagree, and my suspicion is that even you, the reader, will have doubts about my conclusion that scientists sometimes justifi edly infer that a true generalization is caused by nature. Even supposing we know that Nailed to Hume’s Cross? 77 energy is always conserved, there is no getting around the fact that, on my account, it is consistent with all the evidence that our scientists have that this regularity is uncaused (and so a cosmic coincidence), and it is consistent with all that same evi- dence that it is caused by nature (and so a law of nature). There are skeptical arguments that seem to show we don’t know much of anything. Sometimes these take the form of a skeptical-alternative attack. If I am at the zoo in front of a cage labeled ‘Zebra’ and see, standing in front of me, a four-legged striped mammal that I take to be a zebra, a friend might give me pause to think by claiming that it might be a mule disguised to look like a zebra. Since such a mule might look just like the animal before me, it can seem that I do not know the animal before me is a zebra. In pointing out that, even given all the observational evidence, if HS is false, then there might be no explanation of why energy is conserved, Earman and Roberts would be raising a possibility like the possibility of a cleverly disguised mule at the zoo. An epistemological contextualist maintains that the truth-value of an utterance of a sentence containing certain epistemological terms (e.g., the verb ‘to know’ or the adjective ‘justifi ed’) may vary depending on the context.6 So, said in one context, ‘I know that animal is a zebra’ may be true. For example, this might be the case in a discussion between me and a young child who is insisting that the animal is a gazelle. But, said in another context, say one in which I admit that I do not know that the animal is not a cleverly disguised mule, my utterance of that very same sentence would be false. As contextualism is sometimes described, in the fi rst context, the only relevant alternative was that the animal was a gazelle, but, in the second context, there was the relevant alternative that it was a cleverly disguised mule. What is required for a knowledge utterance to be true in a context C is that the cognizer be able to rule out all the alternative hypotheses that are relevant in C. (Keep in mind that in different contexts, different alternatives will be relevant.) This hypothesis about the context sensitivity of epistemological utterances is then used to explain why skeptical arguments of various sorts can seem so convincing and also to mitigate the damage done by those arguments. As the contextualist sees it, even though certain skeptical arguments generate contexts in which many or all knowledge sentences turn out false, this leaves open that there will be lots of important contexts in which utter- ances of those same sentences will turn out true. How does all this apply to the skeptical argument advanced by Earman and Roberts in favor of HS? As Earman and Roberts describe it,
The contextualist maneuver might run as follows: “In contexts where scientists are evaluating a law-positing theory such as T, which is well-supported according to the ordinary standards of scientifi c inference, alternatives such as T* which differ from T only in that they call one or more laws posited by T* nomologically contingent, are not relevant alternatives. Hence, it is not necessary, in order to be justifi ed in believing T, to have evidence that favors T over T *. So Premise 2 of our epistemological argument for HS is false.” (2005: 274)
Now, obviously, Earman and Roberts don’t think that this maneuver will work, but we need not delve into their reasons for thinking so. We need not, because the quote just given misrepresents contextualism. Earman and Roberts bill the contextualist way John W. Carroll 78 out as a method of objecting to premise 2 of their argument. But, in fact, what is crucial to the contextualist reply is that in certain contexts (maybe all contexts!) an utterance of the sentence expressing premise 2 will be true. The antecedent of premise 2 and the statement of premise 1, in virtue of including mention of both T and T*, make (or tends to make) T* a relevant alternative to T, and it is one that cannot be ruled out. To argue as Earman and Roberts do is like arguing: (1) For all we know, given all the evidence presently available to us, nothing favors it being a zebra over a disguised mule and nothing favors it being a disguised mule over a zebra. (2) If nothing favors it being a zebra and nothing favors it being a dis- guised mule then we can’t know it is a zebra. When such premise sentences are uttered and taken seriously by the audience, the contextualist wants to “concede” that then the premise sentences are true and that the conclusion sentence is as well. The con- textualist gambit is now to argue that, once we properly understand the contextual nature of ‘to know’, the fact that an utterance of the conclusion sentence is true is hardly worrisome. This does not rule out that there are other much more ordinary contexts in which an utterance of ‘I know T’ is perfectly true. Science can go on, claims of lawhood are sometimes made, reports of knowledge that such and such is a law sometimes turn out true. The fact that the conclusion sentence of Earman and Roberts’ argument is true, perhaps even as uttered by Earman and Roberts, is no more worrisome than is the fact that utterances of ‘I don’t know the animal is a zebra’ are true in contexts in which the disguised-mule hypothesis is a relevant alternative. Regarding laws, there will be contexts where we presuppose that L is not a coinci- dence. If I have ruled out that there is some in-nature explanation of L, and the presupposition of our conversation is true, then it will be true to say ‘I know T’. Is this enough for science? I think so. According to contextualism, even ‘No one has ever known or will ever know that they have hands’ is true in some contexts, ones in which the evil-demon hypothesis is a relevant alternative. That there are some contexts arising in philosophical discussions in which the sentence ‘No scientist has ever known or will ever know what the laws of nature are’ is true seems mild in comparison. As well it should. That there are contexts where an utterance of ‘I know that I have hands’ is false does not in the least bit undermine the value of my utter- ance of that sentence in certain contexts in which it is true. The contextualist can concede that Earman and Roberts have generated a context in which ‘Scientists know what the laws are’ is false. But, as far as I can tell, that does not generate any absurd or undesirable consequences about science or scientists. What contextualism does is allow us to explain why the Hume-inspired skeptical challenge, at least as raised by Earman and Roberts, can seem so utterly convincing. If the context is right, what they say is convincing because the argument they advance is sound.
5 Avoiding the Cross
I hope that the controversial nature of my replies to the foregoing Hume-inspired metaphysical and skeptical worries provide further evidence that anti-reductionism is not any sort of philosophical dead end. Indeed, the position described here puts in special focus some particularly interesting issues that have the benefi t of potentially Nailed to Hume’s Cross? 79 being more manageable than the search for a reductive analysis of lawhood. First, there is the analysis of coincidence. I have offered what I take to be a plausible account, but coincidence is not a notion that has received anywhere near the atten- tion that lawhood has. Second, the analysis of lawhood offered here depends crucially on the possibility of regularities holding because of nature. While there has certainly been discussion and awareness that regularities are sometimes explained in science, philosophers of science have seemed much more comfortable when the explanandum is some singular or particular fact. My analysis provides new reason to explore the nature of explanations of regularities. Finally, there is the issue in the philosophy of language and linguistics as to how best to describe and understand the context dependence of ‘because’ and ‘knows’. As is the case with many philosophical problems, attention to the language we use to present and address the philosophical problem of laws is a sensible precaution that may help squelch the lure of reductionism.7
Notes
1 Richard Sorabji (1979: 9) attributes this account of coincidence to Aristotle. David Owens (1992: 6) defi nes a coincidence as an event whose constituents are produced by independent causal processes, but maintains that his defi nition has the consequence that all coincidences are inexplicable. 2 This may sound a little odd. My claim that all laws have explanations will strike some as counterintuitive. Aren’t there any laws that are fundamental or basic? Don’t some laws explain though they themselves are unexplained? It is important to keep in mind that what has to be explained in order for P to be a law is P. It is that generalization that cannot be coincidental if it is also to be a law. This leaves open that the lawhood fact, the fact that P is a law, is unexplained; it may be a coincidence. So, for example, it may be a funda- mental law that all inertial bodies have no acceleration, even though something explains why all inertial bodies have no acceleration. It would be a fundamental law in virtue of there being no explanation why it is a law that all inertial bodies have no acceleration. 3 Randy Clarke (2003: 196–217) in his discussion of agent causation surveys the few argu- ments actually given against the possibility of substance causation. 4 See Anscombe (1971), Fales (1990) and Armstrong (1997). 5 This is close to the exact wording of Earman and Roberts’ argument (pp. 257–8). Premise 2 has been simplifi ed by removing words from its antecedent to the effect that realism about lawhood is true. Anti-reductionist anti-realists are not targets of this skeptical attack. 6 There are many versions of contextualism. See, for example, DeRose (1995). My own version is sketched in Carroll (2005). 7 Versions of this paper were presented at Virginia Tech in 2006, at Rutgers University in the spring of 2005 and at NC State University in 2004. Thanks to Troy Cross, Jeff Kasser, Marc Lange, Michael Pendlebury, Jamaal Pitt, Ann Rives, David Robb, John Roberts, and Dean Zimmerman for helpful comments and questions on earlier renditions.
References
Anscombe, G. 1971. Causality and Determination (Cambridge: Cambridge University Press). Armstrong, D. 1983. What Is a Law of Nature? (Cambridge: Cambridge University Press). John W. Carroll 80 ——. 1997. A World of States of Affairs (New York: Cambridge University Press). Blackburn, S. 1984. Spreading the Word (Oxford: Clarendon Press). ——. 1986. “Morals and Modals,” in G. Macdonald and C. Wright, eds., Fact, Science and Moral- ity (Oxford: Basil Blackwell). Carroll, J. 1994. Laws of Nature (Cambridge: Cambridge University Press). ——. 2005. “Boundary in Context,” Acta Analytica 20: 43–54. Clarke, R. 2003. Libertarian Accounts of Free Will (New York: Oxford University Press). DeRose, K. 1995. “Solving the Skeptical Problem,” Philosophical Review 104: 1–52. Dretske, F. 1977. “Laws of Nature,” Philosophy of Science 44: 248–68. Earman, J., and Roberts, J. 2005. “Contact with the Nomic: A Challenge for Deniers of Humean Supervenience about Laws of Nature (Part II),” Philosophy and Phenomenological Research 71: 253–86. Fales, E. 1990. Causation and Universals (London: Routledge). Foster, J. 1983. “Induction, Explanation and Natural Necessity,” Proceedings of the Aristotelian Society 83: 87–101. Giere, R. 1999. Science Without Laws (Chicago: University of Chicago Press). Hume, D. 1955. An Inquiry Concerning Human Understanding (Indianapolis: Bobbs-Merrill). Lange, M. 2000. Natural Laws in Scientifi c Practice (Oxford: Oxford University Press). Lewis, D. 1973. Counterfactuals (Cambridge, MA: Harvard University Press). ——. 1983. “New Work for a Theory of Universals,” Australasian Journal of Philosophy 61: 343–77. ——. 1986. Philosophical Papers, vol. II (New York: Oxford University Press). Owens, D. 1992. Causes and Coincidences (Cambridge: Cambridge University Press). Sorabji, R. 1979. Necessity, Cause and Blame (New York: Cornell University Press). Tooley, M. 1977. “The Nature of Laws,” Canadian Journal of Philosophy 7: 667–98. ——. 1987. Causation (Oxford: Clarendon Press). van Fraassen, B. 1989. Laws and Symmetry (Oxford: Clarendon Press). Ward, B. 2002. “Humeanism Without Humean Supervenience: A Projectivist Account of Laws and Possibilities,” Philosophical Studies 107: 191–218. ——. 2003. “Sometimes the World is not Enough: The Pursuit of Explanatory Laws in a Humean World,” Pacifi c Philosophical Quarterly 84: 175–97.
Nailed to Hume’s Cross? 81 CHAPTER 2.2
Causation and Laws of Nature: Reductionism
Jonathan Schaffer
Causation and the laws of nature are nothing over and above the pattern of events, just like a movie is nothing over and above the sequence of frames. Or so I will argue. The position I will argue for is broadly inspired by Hume and Lewis, and may be expressed in the slogan: what must be, must be grounded in what is. Roadmap: In sections 1 and 2, I will clarify the reductionist thesis, and connect it to a general thesis about modal and occurrent entities. In section 3, I will argue halfway towards reductionism, by arguing that causation reduces to history plus the laws. In section 4, I will complete the case for reductionism, by arguing that the laws themselves reduce to history.
1 Clarifying the Thesis
The reductionist thesis may be expressed as follows:
R1 Causation and laws reduce to history.
But it is not obvious what R1 means, much less why one should believe in it. In this section I will clarify the notions of causation, lawhood, history, and reduction, to the point where arguments may be considered. Starting with causation, the intended notion is perhaps best introduced by exam- ples. Causation is present when one billiard ball strikes another (which Hume called ‘a perfect instance’ of causation), when a person lifts a suitcase, and when a spring uncoils. To a fi rst approximation, one might test for causation by testing for coun- terfactual dependence: if the cause had not occurred, then the effect would not have occurred.1 For present purposes, the notion of causation may be left intuitive. To the extent we understand any notion, we understand this one. Turning to lawhood, what is intended is the sort of thing expressed by the equations of scientists, such as Newton’s laws of motion and Schrödinger’s law of wavefunction evolution. Laws are generally supposed to be expressed by true univer- sally quantifi ed conditionals, holding in a restricted sphere of possible worlds, encoding how the world evolves through time.2 For present purposes, this notion may also be left intuitive. Moving to history, what is intended is the fusion of all events throughout space- time.3 Each individual event is a concrete particular with an intrinsic nature – what occurs in some region of space-time. History is the whole of this – it is what occurs in all of space-time. History is the total pattern of events. Each event is like a bit of a frame in the movie, and history is the whole picture. The notion of reduction is intended to be an ontological relation, expressing depen- dence between entities. As an ontological relation, the intended notion must be distinguished from theoretical and defi nitional relations which may also be labeled ‘reduction’. Theoretical reduction concerns terms found in theories. Defi nitional reduc- tion concerns concepts found in the mind. Ontological reduction is independent of how we conceptualize entities, or theorize about them. Ontological reduction is a thesis about mind-and-theory-independent reality. As a relation of dependence, the intended notion of reduction may be glossed in terms of grounding. What reduces is grounded in, based on, existent in virtue of, and nothing over and above, what it reduces to. What does not reduce is basic, funda- mental, and brute. By way of parable: to create what reduces, God would only need to create what it reduces to. In general, to create the world, God would only need to create what is basic.4 To illustrate, consider the relation between the movie and the frames. Here it is natural to say that the movie depends on the frames. The movie is grounded in, exists in virtue of, and is nothing over and above the frames. To create the movie, the director only needs to arrange the frames. (This is true regardless of how we concep- tualize movies, or theorize about them.) To take a more philosophically interesting illustration, consider the relation between the physical properties, and the mental and moral properties. The physicalist holds that the physical properties are basic, and that the mental and moral properties are grounded in them. According to the physicalist, all God would need to create would be the physical realm.5 R1 is thus the thesis that causation and laws are grounded in, based on, and nothing over and above history, in the way that the movie is grounded in the frames, and the mental (by physicalist lights) is grounded in the physical. To make the world, God only needed to create space-time and fi ll it with intrinsic features. He did not also have to sew the whole thing up with some sort of causal thread. Or so says the friend of R1 – the reductionist. The foe of R1 – the primitivist – holds that a space-time fi lled with intrinsic features still needs sewing together.6 So understood, R1 has implications for explanation and possibility. As to explana- tion, the existence and nature of what reduces is explicable in terms of what it reduces to. What reduces is no further mystery. For instance, the existence and nature of the movie is explicable in terms of the frames.7 Likewise, R1 entails that the existence and nature of causation and lawhood is explicable in terms of history. Causation and Laws of Nature: Reductionism 83 As to possibility, the intended notion of reduction entails that it is impossible for what reduces to differ in any way, without some difference in what it reduces to. In other words, if the reductive grounds are held fi xed, then what reduces must hold fi xed as well. For instance, if the frames are held fi xed, then the movie must hold fi xed as well – the movie cannot differ in any way, without some difference in the frames. Likewise, R1 entails that if history is held fi xed, then the causes and laws must hold fi xed as well. So here are two good ways to try to rebut R1. First, one might argue that history fails to explain causation and laws. Second, one might argue that it is possible for the causes or laws to be different, without any difference in history. If either argument could be substantiated, it would refute R1 – it would prove that causation and the laws fail to reduce, in the intended sense. And here is a bad way to try to rebut R1. One might argue that the concepts of causation and lawhood fail to have conceptual analyses in terms of the concept of history. Because reduction is an ontological relation, and not a relation between concepts in our mind, failure of analysis does not show failure of reduction. There can be reduction without analysis, in the cases where either our concept is insuffi - ciently explicit, or our intuitions are misleading, or the reduction would require an infi nite defi nition.8 This matters, since conceptual analyses of natural concepts virtually always fail. If analysis were required for reduction, then one would likely have to be a primitivist about virtually everything, including movies, marching bands, and motor homes, which do not seem like irreducible elements of reality! So the question of whether causation and the laws of nature reduce to history yields the following questions: (i) are causation and the laws explicable in terms of history? and (ii) is it possible for there to be differences in causation or the laws without any difference in history? Explanation and possibility are reasonably well- understood notions. Thus the question of reduction may prove tractable.
2 Generalizing the Thesis
The reductionist thesis R1 is an instance of a more general thesis about the relation of the modal to the occurrent. Displaying the more general thesis may help to both explicate and motivate R1. The more general thesis is:
R2 Modal entities reduce to occurrent entities.
I will now clarify the notions of the modal and the occurrent, explain how R2 is a generalization of R1, and then provide some motivation for R2. First the clarifi cations: the notion of a modal entity is perhaps best introduced by examples. Paradigmatic modal entities include dispositions, counterfactual properties, and powers. A match, for instance, is disposed to ignite if struck under suitable condi- tions, has the counterfactual property of being such that if it were struck under suitable conditions it would ignite, and has the power to ignite. These are properties of the match, concerning a potentiality the match has (which may not ever be actualized, Jonathan Schaffer 84 if the match is never struck). The match also, in contrast, has the property of having a tip made of potassium chlorate and sesquisulfi de of phosphorus. These are paradig- matic occurrent properties, which simply concern the way the match actually is. It is diffi cult to characterize the distinction between the modal and the occurrent in more precise terms.9 Sider provides the following characterization: “Categorical [/occurrent] properties involve what objects actually are like, whereas hypothetical [/modal] properties ‘point beyond’ their instances” (2001: 41). The sense in which modal properties “point beyond” their instances is that they concern what else must be, while the sense in which occurrent properties remain self-contained is that they concern just the actual, intrinsic features of the thing itself. This characterization could perhaps use further work, but should be suffi ciently clear to work with here. Causation and the laws are clear cases of modal entities. With causation, if c has the property of causing e, then this property of c “points beyond,” adding what else must be the case, namely that e must follow. Causation represents a necessary con- nection between distinct events. Likewise with laws, if c has the property of lawfully entailing e, then this property of c “points beyond,” adding what else must be the case, namely that e must follow. Laws represent necessary connectors between events. Both causation and laws involve not just the actual event itself, but the modal aspect of what else must be. Both causation and laws concern natural necessity. History, in contrast, is a clear case of an occurrent entity. Each individual event is a particular occurrence. It is what is happening in some region of space-time. History is the sum of all the particular occurrences. It is what is happening in all of space-time. Indeed, history exhausts the occurrences. The occurrent aspect of the world is completely given in the pattern of events. This is the whole picture. No occurrences are left out.10 I am now in position to explain how R2 is a generalization of R1. Since causation and the laws are modal entities, R2 entails that causation and the laws reduce to some occurrent entities. Since history exhausts the occurrences, whichever occurrent entities causation and laws reduce to will be found within history. Thus causation and laws reduce to history, as per R1. Recall (section 1) that the reductionist’s general thesis is that God (as it were) only needs to create space-time and fi ll it with intrinsic features. God does not also have to string the whole thing up with modal hooks. The reduction of causation and lawhood to history is just a special case of this thesis.11 It remains to motivate R2. I offer three motivations. First, R2 is intrinsically plau- sible. Modal entities by themselves seem shadowy and mysterious. It seems they cannot fl oat free – they need grounding in the occurrent. What must be, must be grounded in what is. Second, R2 fi ts with a plausible principle about what is possible – Humean recom- bination – which Lewis glosses as: “anything can coexist with anything else, at least provided they occupy distinct spatiotemporal positions” (1986a: 87). Modal entities, insofar as they “point beyond” to what else must be, have implications for what else may exist. Primitive modal entities thus entail implausible limitations on recombina- tion. For instance, if c is accorded the basic property of causing e, then the intuitive possibility of c without e is lost. Causation and Laws of Nature: Reductionism 85 The third motivation for R2 is that it is theoretically useful, in ruling out certain metaphysical views that are now widely regarded as implausible. One example of such an implausible view is Rylean behaviorism, which invokes behavioral dispositions without any occurrent mental grounds. A second example is the view that counter- factuals are primitive. On this view, it might be true of the match that, if it were struck under suitable circumstances, it would ignite, without there being any intrinsic physical features of the match (having phosphorus on the tip) which grounds this counterfactual. A third example is the view that powers are primitive. On this view, it might be true of the match that it has the power to ignite, without there being any intrinsic physical feature of the match which grounds this power. Since dispositions, counterfactuals, and powers are paradigmatic modal entities, the natural generaliza- tion to draw is that primitive modal entities are generally to be shunned.12 The denier of R1 thus faces the question of where to draw the line. If primitive unreduced causal relations or laws are tolerable, what if anything were wrong with Rylean behaviorism, brute counterfactuals, or ungrounded powers? R2 represents a principled and plausible limit on what can serve as a basic feature of reality, which has R1 as a corollary.
3 Defending the Thesis, Stage One: Causation
In what remains I will return to the reductionist thesis R1, and discuss specifi c argu- ments for and against. In this section I will argue halfway towards R1, by arguing that causation reduces to history plus the laws. That is, I will now defend the follow- ing half-reductive thesis:
R3 Causation reduces to history plus laws.
I will fi rst reply to the three main arguments in the literature against R3, and then consider three arguments for R3. I will conclude that the reduction of causation is justifi ed on methodological and scientifi c grounds. Why bother with the half-reductive thesis R3? First, R3 is an important and con- tentious thesis in its own right. Second, R3 enables the reductionist to pursue a divide-and-conquer strategy – all that would remain is to argue that the laws them- selves reduce to history (see section 4). Third, the reductionist may wish to reserve R3 as a fallback position. The case for R3 will prove stronger than the case for the fully reductive R1, because laws have a more secure place in science than causation does (see section 3.2).
3.1 Three arguments for infl ating causation There are three main arguments in the literature against R3, the fi rst of which is that the concept of causation cannot be analyzed via history plus the laws. An analysis is an attempt at providing fi nite, non-circular, and intuitively adequate necessary and suffi cient conditions. So an analysis of causation via history plus laws will look something like: c causes e iff ——, where the blank is to be fi lled in with some fi nite Jonathan Schaffer 86 entry concerning history and laws (and not containing the term ‘causation’ or any other terms that are themselves to be analyzed via causation), and where the resulting entry will match our intuitive judgments about whether causation obtains in most if not all conceivable cases. So the fi rst argument is that it is impossible (or at least unlikely) that the schema ‘c causes e iff ——’ can be completed in this way. The argument from unanalyzability might be formulated as follows:
(1) The concept of causation cannot be analyzed in terms of history plus laws. (2) If the concept of causation cannot be analyzed in terms of history plus laws, then causation does not reduce to history plus laws. (3) Causation does not reduce to history plus laws.
The argument is valid, so the only question is of the truth of premises 1 and 2. I would accept (1). There is a long history of attempts to analyze the concept of causation in various terms.13 All the attempts have been riddled by counterexamples (as have attempts at conceptual analyses for all natural concepts – there is nothing special about causation here). Obviously this does not prove that no analysis is pos- sible. But the prospects seem bleak. What I would deny is (2). (2) seems to be presupposed by most who have attempted conceptual analyses – this is why they have attempted it.14 But ontological reduction does not require conceptual analysis (section 1). Perhaps our concept is insuffi ciently explicit, or our intuitions are misleading, or the analysis would require an infi nite defi nition. From a failure of conceptual analysis, nothing follows about the world. The philosopher who would uphold the unanalyzability argument must explain (i) how the concept of causation differs from any other natural concept and (ii) why this conceptual difference is relevant to how the world is. The best attempt I know of to maintain that the concept of causation differs from other natural concepts is via the claim that the causal concept is especially central to our conceptual scheme.15 Perhaps so. But even so, such a conceptual difference does not seem relevant to how the world is. One could imagine a creature wired to think of everything through the concepts of edible and inedible – what would that prove about the world? So I con- clude that the fi rst argument confuses a conceptual with an ontological issue. The second main argument in the literature against R3 is that events themselves can only be individuated causally. Individuation principles are attempts to describe how to count entities in a given domain, by saying when there is one. So individua- tion principles for events will look something like: there are no two events e1 and e2 (where e1 ≠ e2) such that ——, where the blank is to be fi lled in by whatever factors would render e1 and e2 one and the same. So the second argument is that it is impos- sible to individuate events except in terms of their causes and effects. For instance, one might hold that the relevant individuation condition is: there are no two events e1 and e2 (where e1 ≠ e2) such that e1 and e2 differ in their causes and effects.16 The argument from event individuation might be formulated as follows:
(4) Events can only be individuated in causal terms. (5) If events can only be individuated in causal terms, then causation does not reduce to history plus laws. (6) Causation does not reduce to history plus laws. Causation and Laws of Nature: Reductionism 87 The argument is valid, so the only question is of the truth of (4) and (5). I would accept (5). If events can only be individuated in causal terms, then (onto- logically speaking) there can be no history that is prior to causation. For history itself, as the sum of all actual events, would presuppose some defi nite number of events in defi nite relations, which would itself depend on causation. God could not, as it were, simply intone: “Let there be history!” for God would fi rst need to create the causal relations that shape history. What I would deny is (4). Events can be individuated in purely occurrent terms, by their spatiotemporal locations and intrinsic natures. That is, I would offer the fol- lowing non-causal individuation principle: there are no two events e1 and e2 (where e1 ≠ e2) such that e1 and e2 occupy the same spatiotemporal region and possess the same intrinsic nature.17 For instance, if e1 is an instance of red here, and e1 ≠ e2, then e2 cannot also be an instance of red here – e2 must concern some other property or some other region. Those who uphold (4) typically argue that the postulation of intrinsic natures is epistemically disastrous. They argue that our only epistemic access to events is through their effects (specifi cally, their effects on our minds), so to postulate natures is to suffer a skeptical fate. Such natures could never be known save through their effects, so the ontologist should just drop the natures and limit herself to the effects.18 My reply to the skeptical concern is that it embodies disastrous epistemic reason- ing. For the same sort of reasoning should lead us to drop the effects and skip straight to the effects on us (for our only epistemic access to the effects is through their effects on us). This would be to drop external reality and only recognize the contents of one’s own mind – this would be solipsism. Contrapositively, once we countenance an external reality (and who would reject that?) we are already dabbling in entities we cannot directly access. Intrinsic natures of events are just more of the same.19 So I conclude that the second argument makes unsustainable epistemic assumptions. The third main argument in the literature against R3 is that it is possible for worlds to differ in causation without differing in history or laws. To my mind, this is the most serious argument against causal reduction. Here is a representative example. Imagine that two wizards, Merlin and Morgana, each cast a spell to turn the prince into a frog, and that the prince then transforms into a frog. Imagine that all spells have a 50 percent chance of success, according to the laws of this fantasy world. Now, the argument proceeds, intuitively there are at least three distinct possibilities. First, it is possible that Merlin’s spell alone caused the transformation. Second, it is possible that Morgana’s spell alone caused the transformation. Third, it is possible that both spells causally overdetermined the transformation. These three distinct pos- sibilities involve the same histories and the same laws. So the argument concludes that there can be differences in causation without differences in history or the laws. If correct, this would refute R3.20 The argument from causal differences might be formulated as follows:
(7) There are worlds that differ in causation without differing in history or laws. (8) If there are worlds that differ in causation without differing in history or laws, then causation does not reduce to history plus laws. (9) Causation does not reduce to history plus laws. Jonathan Schaffer 88 The argument is valid. But are (7) and (8) true? I would accept (8). Ontological reduction has implications for possibility (section 1), such that if there really are two possible worlds that differ in causation without differing in history or the laws, then R3 would stand refuted. What I would deny is (7). Why believe that there are genuinely distinct possibilities here? To my mind there is only the one possibility (the one in which Merlin and Morgana both cast their spells, and the prince transforms), confusingly described in three different styles. For in what respect are these alleged possibilities said to differ? The alleged causal difference seems to fl oat on nothing – it seems a verbal distinction without any genuine ontological difference.21 The philosopher who would uphold (7) would presumably reply that the reason for accepting genuinely distinct possibilities here is intuitive, and that this shows that the alleged causal difference need not rest on anything – it is a brute and fundamental difference. But this seems a terrible metaphysical price for a relatively fl imsy intuition (section 2). Or at least, staying within the realm of intuitions, I would say that I have strong countervailing intuitions that causal facts (and modal facts generally) cannot fl oat free like this. So at most the intuitive argument for (7) has revealed confl icted intuitions, rather than a clear stance against R3. So the question of whether (7) is true yields the question of how to weigh confl ict- ing intuitions. I do not have a general answer to this question. But it seems to me that the reductionist can explain away the primitivist intuitions, from the conceptual error of reifi cation. Reifi cation occurs when a mere concept is mistaken for a thing. We seem generally prone to this sort of error. Our causal vocabulary allows us three different descriptions, and this leaves us prone to positing three different possibilities. So I would conclude with the suggestion that the third and most serious argument against R3 trades on fl imsy reifi cations.22
3.2 Three arguments for reducing causation Here are three main arguments for R3, the fi rst of which is that causal knowledge requires reduction. The idea is that our causal knowledge is ultimately based on our observation of regularities in history, so that if there were more to causation than such regularities, we could have no access to this further feature. Such a feature could not be discovered save through the regularities it engenders, so the ontologist should just drop the further feature and limit herself to the regularities.23 The argument from causal knowledge might be formulated as follows:
(10) We possess causal knowledge. (11) If we possess causal knowledge, then causation must reduce to history plus laws. (12) Causation must reduce to history plus laws.
The argument is valid, and I take it no one would deny (10).24 So the only serious question is the truth of (11). Though I favor reduction, I would deny (11). Presumably (11) might be defended as follows: Causation and Laws of Nature: Reductionism 89 (11a) If we possess causal knowledge, then causation must be nothing over and above what is directly observable. (11b) All that is directly observable is history. (11c) If we possess causal knowledge, then causation must be nothing over and above history.
Here the idea is that anything beyond the actual pattern of events would escape our direct observation and thus escape our knowledge. As such, this is but another falla- cious leap from knowledge to direct access. That is, once we allow that knowledge is possible without direct access (as with knowledge of the external world: section 3.1), then we must either deny (11a) or liberalize what counts as ‘directly observable’ so as to deny (11b). The way to deny (11a) would be to allow that we can fi nd indirect theoretical warrant for causation, at least in favorable cases. Here the idea is that (i) causal rela- tions have directly observable consequences, such that (ii) directly observing such consequences furnishes abductive evidence for postulating causal relations. More formally, all that is required is that one might rationally set one’s credences such that some bit of evidence E is taken to raise the probability of some hypothesis H. Then discovering E will furnish evidence for H. Here E may be a directly observable fact, and H a causal hypothesis.25 The way to deny (11b) would be to allow that we can directly observe some causal relations, in the requisite sense. For instance, one might argue that we can directly observe the causal relation in certain very special cases, such as between willing and action, and/or between pressure on the body and the sensation of it.26 Or one might argue that we can directly observe the causal relation in a wide range of cases, such as when we see the boulder fl atten the hut, or when we see the man pick up the suitcase and lift it on to the rack. Is this not seeing causation?27 I take no stand on whether the infl ationist should deny (11a) or (11b) (this question involves diffi cult issues concerning perception). But one way or another, I would deny (11). The epistemic reasoning behind (11) seems to be of a disastrously skeptical sort. If one holds that all that is ultimately directly observable are sense-data, then parallel reasoning will force one to solipsism. If one allows that parts of the external world are directly observable, then no special reason has been given to resist direct causal knowledge. What is interesting here is that the epistemic reasoning behind (11) seems of a piece with the epistemic reasoning behind (4) (which led to the rejection of intrinsic natures for events). So one should conclude that if such epistemic reasoning is acceptable, then both the reductionist and the non-reductionist are in trouble, as the former posits inaccessible causation and the latter posits inaccessible natures. In the other direction, if one thinks that either the reductionist or the non-reductionist is right, one must have equal disdain for both (4) and (11). So I must conclude that the most historically important argument for R3 embodies disastrous epistemic reasoning.28 The second main argument for R3 is that sound methodological principles support reduction. To posit an irreducible causal reduction is to offend against (i) theoretical fathomability, and (ii) ontological economy. The argument from theoretical fathom- ability proceeds by pointing out that necessary connections have an air of the occult, Jonathan Schaffer 90 implying inexplicable necessary connections between distinct existents (section 2). The argument from economy proceeds by invoking Ockham’s Razor: one should not multiply entities beyond necessity. It is then maintained that it is not necessary to introduce irreducible causal relations. Or at least, none of the arguments canvassed above (section 3.1) show any deep necessity. The argument from methodology might be formulated as follows:
(13) Sound methodological principles (such as theoretical fathomability and onto- logical economy) support reducing causation to history plus laws. (14) If sound methodological principles support reducing causation to history plus laws, then causation must reduce to history plus laws. (15) Causation must reduce to history plus laws.
The argument is valid but unsound, since (14) is clearly false. Merely methodological principles can be outweighed (for instance, Ockham’s Razor only tells us not to pos- tulate entities without necessity). They are merely prima facie constraints. A more nuanced formulation would replace (14) and (15) with:
(14′) If sound methodological principles support reducing causation to history plus the laws, then, unless suffi ciently countervailing considerations can be adduced, causation must reduce to history plus laws. (15′) Unless suffi ciently countervailing considerations can be adduced, causation must reduce to history plus laws.
I take it no one would deny the resulting argument ((13), (14′), and (15′)).29 But now the conclusion is hedged, and the question of reduction is just the question of whether there are suffi ciently countervailing considerations to be adduced. For that is what it would take to move from the uncontentious (15′) back to the contentious (15). The infl ationist must now return to her arguments (c.f. the three main arguments of section 3.1), to identify suffi ciently countervailing considerations. I see none. The argument from unanalyzability seems to me to be a mere confusion between concep- tual primitiveness and ontological primitiveness. The argument from event individu- ation seems to me to be a mere invocation of otherwise disastrous epistemic reasoning (which would doom infl ationism anyway via the argument for causal knowledge). And the argument from causal differences seems to me to be a mere exercise in rei- fi cation. But here there is a further point to be made, which is that even if there is some residual intuitiveness to the argument from causal differences, surely it is not suffi ciently powerful to overturn the push for a fathomable and economical theory. After all, such a highly questionable intuition hardly seems suffi cient to generate the sort of necessity needed to blunt Ockham’s Razor. So I conclude that the methodologi- cal argument for R3 is ultima facie successful, even granting some intuitiveness to the infl ationist arguments. The third main argument for R3 is that scientifi c practice supports reduction. To my mind, this is the best argument for causal reductionism. The idea is that sophisti- cated science invokes only laws and events. Causation drops out as an imprecise, folk mode of description. So it is concluded that causal relations, if they are real at all, Causation and Laws of Nature: Reductionism 91 must be nothing over and above the laws and events that serious scientifi c practice requires. The argument from scientifi c practice might be formulated as follows: (16) Scientifi c practice only requires history and laws. (17) If scientifi c practice only requires history and laws, then causation must reduce to history plus laws. (18) Causation must reduce to history plus laws. The argument is valid, so it remains to ask if (16) and (17) are true. The case for (16) is that causation disappears from sophisticated physics. What one fi nds instead are differential equations (mathematical formulae expressing laws of temporal evolution). These equations make no mention of causation.30 Of course, sci- entists may continue to speak in causal terms when popularizing their results, but the results themselves – the serious business of science – are generated independently. There are two main ways that the infl ationist might oppose (16), the fi rst of which is to maintain that causation is still integral to the practice of the special sciences. Here it would be argued that (i) special sciences – especially the social sciences – remain suffused with causal notions, such that (ii) (16) is false when sciences other than physics are considered. (Here the infl ationist might rail against special privileges for physics.) The problem with this fi rst sort of reply is that it would amount to maintaining that there is a brute and fundamental feature of reality that is only accessible to the special sciences. It would mean that physicists could in principle answer every question save for what causes what, at which point they would need to consult the economists. That constitutes a reductio. The second main way that the infl ationist might oppose (16) is to maintain that causation is still integral even within physics. Here the most plausible candidate role for causation is in interpreting the relativistic prohibition against superluminal velocities, as a prohibition against superluminal signaling.31 The problem with this second sort of reply is that it seems to presuppose reduction- ism. Invoking causation in the foundations of special relativity is only helpful on the presupposition that certain worldlines (e.g., the billiard ball) are causal processes, while certain worldlines (e.g., Salmon’s spot of light) are non-causal pseudo-processes. But if the occurrence of causation is brute, there is no basis for this presupposition. Only the reductionist can render causation fi t to play a role in the foundations of special relativity. So I conclude that (16) should stand. The case for (17) is that science represents out best attempt at a systematic under- standing of the world, and if a certain notion proves unneeded in our best attempt at a systematic understanding of the world, this provides strong evidence that what this notion concerns is not ontologically basic.32 Of course, one might deny (17) by insisting that there could be more on heaven and earth than is dreamt of in the sciences. Perhaps so. Perhaps there really are witches, vital forces, real simultaneity relations, and other sundries that science has learnt to discard. But I doubt it. Causal relations must either reduce to what is required by science, or else be eliminated.33 So I conclude that (17) should stand, and that the scientifi c argument for R3 succeeds. Causation must reduce, or face elimination. Jonathan Schaffer 92 In this vein it is worth returning again to the argument from causal differences (section 3.1). For we can now add that, even if there is some residual intuitiveness to the argument, and even if such intuitiveness were not immediately trumped by methodological considerations, such an intuition should be dismissed as pre-scientifi c. It is just the afterglow of our ignorance.
3.3 Conclusions on causation So far I have defended the half-reductive thesis R3. I have examined three arguments against R3 and found them wanting, and examined three arguments for R3 and found the methodological and scientifi c arguments successful. In short, infl ated causation represents an unwarranted reifi cation of a folk concept, which is methodologically and scientifi cally suspect. I have not said how causation reduces. That is, I have not said what aspect of history plus laws grounds causal facts. Here I am partial to Lewis’s (1986b) claim that causation has to do with patterns of counterfactual entailments, which are themselves grounded in history. But I am not offering an analysis. I am only treating this as a useful gloss, whose role is to show why there is no further mystery here. (Compare: to see that the movie reduces to the sequence of frames – to see that there is no further mystery there – it suffi ces to have a rough sense of how the sequence of frames comprises the movie. One does not need a conceptual analysis of “movie” for that.) Of course, R3 is only a half-reductive thesis. The infl ationist might accept R3, and simply add that laws are ontologically primitive.34 So it remains to discuss the second half of the reductive thesis R1, which is that laws reduce to history.
4 Defending the Thesis, Stage Two: Laws
To complete the case for reductionism, it remains to argue that laws reduce to history, as per:
R4 Laws reduce to history.
For given R3 and R4, the fully reductive thesis R1 follows – if causation reduces to history plus laws, and laws themselves reduce to history, then both causation and laws must reduce to history. In what remains, I will fi rst reply to the three main arguments in the literature against R4, and then consider three arguments for R4. I will conclude that the reduc- tion of laws is justifi ed on methodological and metaphysical grounds.
4.1 Three arguments for infl ating laws There are three main arguments in the literature against R4, the fi rst of which is that the concept of lawhood cannot be analyzed in terms of history. The idea is that the schema: ‘L is a law of nature iff ——’ cannot be completed in the requisite way (by Causation and Laws of Nature: Reductionism 93 fi nite, non-circular, and intuitively adequate necessary and suffi cient conditions con- cerning history). This argument, which parallels the unanalyzability argument for causation of (1)–(3), might be formulated as follows:
(19) The concept of lawhood cannot be analyzed in terms of history. (20) If the concept of lawhood cannot be analyzed in terms of history, then laws do not reduce to history. (21) Laws do not reduce to history.
My reply to (19)–(21) parallels my reply to (1)–(3). I would accept (19), on grounds of the history of failed attempts at a conceptual analysis of lawhood, and also on grounds of the conceptual centrality of lawhood.35 But I would reject (20), on grounds that a failure of conceptual analysis tells us nothing about the world. I have nothing further to add to what has been said already (section 1), so I will leave the argument here. The second main argument in the literature against R4 is that events can only be individuated in nomic terms. For instance, one might hold that the relevant individu- ation condition is: there are no two events e1 and e2 (where e1 ≠ e2) such that e1 and e2 differ in their nomic relations. This argument, which parallels the individuation argument for causation of (4)–(6), might be formulated as follows:
(22) Events can only be individuated in nomic terms. (23) If events can only be individuated in nomic terms, then laws do not reduce to history. (24) Laws do not reduce to history.
My reply to (22)–(24) parallels my reply to (4)–(6). I would accept (23), since if events can only be individuated in nomic terms, then (ontologically speaking) there can be no history that is prior to the laws. For the pattern of events that is history would itself presuppose some defi nite number of events in defi nite relations, which would itself depend on the laws. But I would reject (22), on grounds that events may be individuated by spatiotemporal locations and intrinsic natures (without untoward skeptical consequences). Here I have nothing further to add to the previous discussion (section 3.1), so I will move forward. The third main argument in the literature against R4 is that it is possible for worlds to differ in laws without differing in history. To my mind this is the most serious argument against nomic reduction. Here is a representative example. Imagine a rela- tively simple world with just a single electron moving in a straight line forever. Now, the argument proceeds, there are (infi nitely) many distinct possibilities. For instance, there are the possibilities that (i) this world is governed by Newton’s three laws; (ii) this world is governed by the single law that all things move in straight lines forever; and (iii) this world is governed by the two laws that all electrons move in straight lines forever, and that all protons spin in mile-radius circles forever. So the argument concludes that there can be differences in lawhood without differences in history. If correct, this would refute R4.36 The argument from nomic differences might be formulated as follows: Jonathan Schaffer 94 (25) There are worlds that differ in lawhood without differing in history. (26) If there are worlds that differ in lawhood without differing in history, then laws do not reduce to history. (27) Laws do not reduce to history.
It remains to ask after (25) and (26). I would accept (26), given the implications that ontological reduction has for possibility (section 1).37 So it remains to ask after (25). I would, of course, deny (25). Why believe that there are (infi nitely many) genuinely distinct possibilities here? To my mind, there is only the one possibility (the one in which a single electron moves in a straight line forever), confusingly described in three different styles. For in what respect are these alleged possibilities said to differ? The alleged nomic difference seems to fl oat on nothing – it seems a verbal distinction without any genuine ontological difference. The case may be even clearer with respect to the empty world, where nothing at all happens. The infl ationist, to my mind, faces an embarrassment of riches here, for she is committed to infi nitely many empty worlds governed by infi nitely many sets of purely vacuous laws. This seems a groundless multiplication.38 The philosopher who would uphold (25) might reply in three main ways, the fi rst of which would be that the reason for accepting genuinely distinct possibilities here is intuitive, where this shows that the alleged nomic difference need not rest on any- thing – the nomic difference is brute. But this seems a terrible metaphysical price for an especially fl imsy intuition (section 2).39 Or at least, staying within the realm of intuitions, I have strong countervailing intuitions that nomic facts (and modal facts generally) cannot fl oat free like this. So at most the intuitive argument for (25) has revealed confl icted intuitions, rather than a clear stance against R4. In any case, there are reasons to be skeptical of the intuitions behind (25). For the notion of lawhood in use is a direct descendant of the theological views of Descartes, Newton, and Leibniz, who viewed laws as divine decrees concerning the clockwork of the world.40 The idea of laws as divine decrees seems to engender the intuitions of distinct possibilities. Here one is intuiting God acting in different ways. But if one rejects the view of laws as divine decrees, it is not clear why one should continue to hold onto the intuitions it engenders. (In particular, if one reinterprets laws as sum- maries of history (section 4.3) then it is clear one should reject these intuitions as misguided.) So I conclude that there is good reason to reject the intuitions involved, as remnants of a dubious theology.41 The second (and perhaps better) defense of (25) would invoke scientifi c practice. Here the idea is that (i) scientists treat, e.g., the empty world as a model of Newtonian mechanics and of other nomic systems, and (ii) what scientists treat as a model of a system of laws should be treated as compossible with those laws.42 But it is unclear that (i) is essential to scientifi c practice, and it seems that (ii) is an additional philo- sophical inference. To reject (ii) is not to reject the practice of modeling, it is only to allow that some models (though they may be useful and interesting) are still models of metaphysically impossible situations. For instance, it might be useful and interest-
ing for a geologist to explore a model in which water is H2SO4, even though such is
metaphysically impossible (necessarily, water is H2O). So the scientifi c practice argu- ment for (25) seems to fall short.43 Causation and Laws of Nature: Reductionism 95 The third (and perhaps best) defense of (25) would involve considerations about counterfactuals. Here the idea is to begin with two more complex worlds. For instance, let w1 be a world in which an electron moves in a straight line forever, and a proton – which comes into existence by chance – does the same. And let w1 have the law that all things move in straight lines forever. Now let w2 be a world in which an electron moves in a straight line forever, and a proton – which also comes into exis- tence by chance – spins in a mile-radius circle forever. And let w2 have the two laws that all electrons move in straight lines forever, and that all protons spin in mile- radius circles forever. What is important about w1 and w2 is that these worlds beg no questions against the reductionist. For the nomic difference between w1 and w2 seems suitably grounded in the different histories at these worlds.44 Now we add the following principle about laws under counterfactuals:
LUC If it is nomologically possible that p, and nomologically necessary that q, then had p been the case, then it would (still) be nomologically necessary that q.
Now from w1 and LUC (given that the existence of the proton is chancy and so might not have obtained) we get the possibility of a single electron world w3 that still has the single straight-line law, while from w2 and LUC we get the distinct possibility of a single electron world w4 that still has the dual electron-line and proton- circle laws.45 But it is unclear why LUC should be endorsed. Whether LUC is valid depends on the correct modal logic for lawhood. In particular, LUC is only valid in K4 or stronger modal systems with transitive accessibility.46 Here is an argument against transitive accessibility for lawhood. Transitive accessibility would function as a meta-law gen- erator. Transitivity entails that if it is a law that p, then it is a law that it is a law that p. Repeated applications entail an infi nite hierarchy of meta-laws: ᮀᮀp, ᮀᮀᮀp, ᮀᮀᮀᮀp, etc. This would mean that the existence of laws entails the existence of laws of laws, laws of the laws of laws, etc. But clearly there is no reason to believe that laws require laws of laws, much less that laws require an infi nite hierarchy of meta- laws. It seems enough just to have the laws. Meta-laws are strange entities, scientifi c practice does not require them, and philosophers have hitherto not dreamt of them. So there is good reason to reject the counterfactual argument for (25), as presupposing a poor modal logic for lawhood.47 And thus I conclude that the third and most serious argument against R4 fails.
4.2 Three arguments for reducing laws Here are three main arguments for R4, the fi rst of which is that nomic knowledge requires reduction. The idea is that our nomic knowledge is ultimately based on our observation of regularities in history, so that if laws were more than such regularities, we could have no access to this further feature. So the ontologist should just drop the further feature and limit herself to the regularities.48 The argument from nomic knowledge might be formulated as follows: Jonathan Schaffer 96 (28) We possess nomic knowledge. (29) If we possess nomic knowledge, then laws reduce to history. (30) Laws reduce to history.
I would reject (29), however, for reasons parallel to my rejection of (4) and (11). There is some difference between the causal and nomic cases, in that it is much less plausible that nomic knowledge can be attained by direct observation. Or at least, the sorts of cases in which one might argue that causation is directly observable (such as opera- tions of the will, seeing the boulder fl atten the hut: section 3.2), do not seem like cases in which the relevant laws (psychological laws and laws of motion, presumably) are themselves directly observable. But this is only to show that lawhood seems more towards the theoretical side of the blurry line between observation and theory. It remains perfectly appropriate, as far as I can see, for the infl ationist to argue that we can directly observe certain sequence of events, that provide evidence for theoretical claims about the laws. So I will not press the argument further. The second main argument for R4 is that sound methodological principles support reduction. To posit an irreducible nomic relation is to offend against theoretical fath- omability and ontological economy. The argument from theoretical fathomability proceeds by pointing out that necessary connections have an air of the occult, in the sense that they imply inexplicable necessary connections between distinct existents (section 2).49 The argument from economy proceeds by invoking Ockham’s Razor: one should not multiply entities beyond necessity. It is then maintained that it is not necessary to introduce irreducible nomic relations. Or at least, none of the arguments canvassed above (section 4.1) shows any real necessity for irreducible laws. The argument from methodology (nuanced to allow for methodological concerns to be overridden, as per section 3.2) might be formulated as follows:
(31) Sound methodological principles (such as theoretical fathomability and onto- logical economy) support reducing laws to history. (32) If sound methodological principles support reducing lawhood to history, then, unless suffi ciently countervailing considerations can be adduced, laws reduce to history. (33) Unless suffi ciently countervailing considerations can be adduced, laws reduce to history.
Here the main question is whether suffi ciently countervailing considerations can be adduced, to discharge the “unless . . .” qualifi cation on (33). The infl ationist must now return to her arguments (section 4.1), to identify suffi - ciently countervailing considerations. I see none. The argument from unanalyzability seems to confuse conceptual primitiveness and ontological primitiveness (an instance of reifi cation). The argument from event individuation seems to me to embody disas- trous epistemic assumptions (which would doom infl ationism anyway via the argument for nomic knowledge). And the argument from nomic differences seems to me steeped in dubious theology. But here there is a further point to be made, which is that even if there is some residual intuitiveness to the argument from nomic dif- ferences, surely it is not suffi ciently powerful to overturn the push for a fathomable Causation and Laws of Nature: Reductionism 97 and economical theory. Such a highly questionable intuition hardly seems suffi cient to generate the sort of necessity needed to blunt Ockham’s Razor. So I conclude that the methodological argument for R4 is ultima facie successful, even granting some intuitiveness to the infl ationist arguments. There is some difference between the methodological arguments for causal reduc- tion (section 3.2) and for nomic reduction, in that irreducible laws seem far less fathomable than irreducible causation. This may be due to the more theoretical, less observable nature of lawhood. In any case, it renders the nomic infl ationist in the perilous position of presenting a completely unfathomable theory, that for us could contain little more than theistic metaphors.50 The third main argument for R4 returns to the need to ground modal entities in occurrent entities, as per R2 (section 2). To my mind this is the best argument for nomic reductionism. Here the idea is that laws are modal, history exhausts occurrent existence, and as such the laws need to be grounded in history. The argument from grounding might be formulated as follows:
(34) Modal existents reduce to occurrent existents. (35) If modal existents reduce to occurrent existents, then laws reduce to history. (36) Laws reduce to history.
But is the argument sound? Are (34) and (35) true? (34) has been defended above (section 2), as intuitively plausible, consistent with modal recombination, and useful in ruling out some bad metaphysics. Laws require grounding. Here the infl ationist must identify some sort of error driving the arguments for (34) (beyond complaining that (34) confl icts with her theory). Pending such a response, I conclude that (34) should stand. The case for (35) has two parts. The fi rst part is the claim that laws are modal existents (section 2). This should be uncontroversial. The second part is the claim that the pattern of events exhausts occurrent existence (for then whatever occurrent existents the laws reduce to will be present in the pattern of events). It is this second part of (35) that proves controversial, for there is a certain sort of infl ationist who would reject it. The sort of infl ationist who would reject the second part of (35) would expand the realm of occurrent existence to include further occurrences alongside history. She would accept the reduction of laws to the occurrent, while denying the reduction of laws to history alone.51 As such, this occurrent infl ationist might seem to respect both the intuition of nomic differences in 25, and the intuition of grounding in (34). This seems like the best of both worlds. So the occurrent infl ationist might claim victory, at this late hour. But the occurrent infl ationist faces an underlying dilemma. The underlying dilemma concerns the modal status of the link between her extra occurrents and the regulari- ties. Never mind what the nature of this link is. Instead, ask of this mysterious link whether it holds necessarily or contingently. For instance, does N(F,G) necessarily entail (∀x) (Fx → Gx), or does N(F,G) only contingently entail (∀x) (Fx → Gx)? In other words, are all the worlds in which N(F,G) exists worlds in which (∀x) (Fx → Gx) holds, or only some? Jonathan Schaffer 98 If the link is said to hold necessarily, then the allegedly “occurrent” lawmaker is revealed as modal after all. For it will concern what else must be the case, namely the regularity. In other words, it will limit how things may be combined. The state Fa and the lawmaker N(F,G) cannot be combined with the absence of Ga. So on this horn, the grounding intuition remains unsatisfi ed. On this horn, the occurrent infl a- tionist fares no better than the nomic primitivist with respect to grounding modal entities in occurrent entities. On the other horn of the dilemma, if the link is said to hold contingently, then the alleged “lawmaker” is revealed as insuffi cient. It will not govern the events. There will be worlds in which the lawmaker exists but the events pay it no heed. For instance, there will be worlds in which N(F,G) exists and Fa obtains, but Ga does not. Further, the nomic differences intuition will remain unsatisfi ed. For there will be worlds that agree on both history and the “lawmakers,” but differ on their linkage. For instance, there will be a world w1 in which N(F,G) exists, (∀x) (Fx → Gx) holds, and these are linked; and there will be a world w2 in which these are unlinked. So on this horn, the occurrent infl ationist fares worse than the reductionist with respect to the platitude that the laws cannot be violated, while faring no better than the reductionist with respect to nomic differences.52 Thus (35) should stand. Thus lawhood must reduce, or lose grounding. In this vein it is worth returning again to the argument from nomic differences (section 4.1). For we can now add that, even if there is some residual intuitiveness to the argument, and even if such intuitiveness were not immediately trumped by meth- odological considerations, such an intuition should be dismissed for groundlessness.
4.3 Conclusions on laws I have now defended the second stage of the reductionist thesis R1 by defending the reduction of laws to history as per R4. I have examined the arguments against R4 and found them wanting, and examined the arguments for R4 and found the meth- odological and metaphysical arguments successful. In short, infl ated lawhood is a methodologically and metaphysically suspect vestige of dubious theology. I have not said how lawhood reduces. That is, I have not said what aspect of history grounds nomic facts. Here I am partial to Lewis’s (1973) suggestion that the laws represent the theorems of the best deductive systematization of the occurrent facts.53 The suggestion is vague (what makes a system “best”?), and I am not interested in providing an analysis. I am only treating this as a useful gloss, whose role is to show why there is no further mystery in lawhood, given the pattern of events. The reduction of lawhood via R4, together with the reduction of causation via R3, entails the reductionist thesis R1. So I am now in a position to conclude that causa- tion and laws reduce to the pattern of events, just like a movie reduces to the sequence of frames. Each event is like a bit of a frame of the movie. There are causes and laws in the movie, but only, as it were, as themes of the plot, present in virtue of what is in the frames. Interesting philosophical disputes arise from confl icted intuitions. The dispute over whether causation and laws reduce arises for this reason. On the one hand, we have infl ationist intuitions about possible causal and nomic differences given the same Causation and Laws of Nature: Reductionism 99 history; while, on the other, we have reductionist intuitions about ontological economy and the metaphysical need for occurrent grounding. Something must give. I have cast aspersions on the infl ationist intuitions, arguing that our intuitions about possible causal differences are due to reifi cations of a folk concept, and that our intuitions about possible nomic differences are due to vestiges of a theological world-view. Perhaps the infl ationist can cast similar aspersions on ontological economy and metaphysical grounding. That is what it would take to counter the line of argument here. Pending such a response, I must conclude that the reductionist view is overall best. To summarize, the reductionist offers a more fathomable and economical theory that respects the need for grounding, while the infl ationist relies on dubious folk and theological intuitions to attempt to convince us that more things exist than we may fathom or need.54
Notes
1 For a sophisticated development of this idea, see Lewis (1986b). Note that I am not sug- gesting that causation may be analyzed in terms of counterfactual dependence, or even suggesting that causation may be analyzed at all. As I will explain below, part of my purpose is to separate the question of ontological reduction from the question of concep- tual analyzability. 2 See Armstrong (1983) for a sophisticated (though anti-reductionist) view of laws along these lines. Note that I am not offering an analysis of lawhood, but only trying to convey the intended notion. 3 Point of clarifi cation: as I use the notion, history includes past, present, and future. It is not limited to the past, to the environs of the earth, or any proper part of the actual world. Though it is limited to the actual world. 4 Here I am following Fine, who suggests: “Reduction is to be understood in terms of fundamental reality” (2001: 26). For further discussion of ontological dependence and basicness, see Fine (1994), Lowe (2005), and Schaffer (forthcoming). 5 Thus Loewer expresses the physicalist credo as follows: “The fundamental properties and facts are physical and everything else obtains in virtue of them” (2001: 39). 6 There is a second sort of foe of R1 – the eliminativist – who denies that there is any cau- sation or lawhood in the world at all. See Russell (1992) for a defense of eliminativism for causation, and van Fraassen (1989) for eliminativism about laws (or at least for “a programme for epistemology and for philosophy of science which will allow them to fl ourish in the absence of laws or belief therein” (1989: 130)). In the main text I will simply be presupposing that causation and lawhood are real. The prospects for eliminativism will not be considered further here. 7 This is not to suppose that the explanation can be written in fi nite terms, or grasped by human minds. If there could be an endless and patternless movie, for instance, we might never succeed in saying how the movie is grounded in its infi nite frames. But it would still be the case that the endless and patternless movie harbors no further mystery (beyond its infi nity of frames). 8 There can also be analysis without reduction, in the cases where either our concept is overly defl ated, or the terms in the defi nition denote entities that actually reduce to those denoted by the target concept, rather than the other way around. The moral of all this is that one must be careful to distinguish the conceptual order, which is an ordering of con- cepts in our minds by the relations “fi gures in the defi nition of,” from the ontological Jonathan Schaffer 100 order, which is an ordering of entities in the world by the reductive relation. Even if there is a conceptual order in the intended sense, it need not track the ontological order in any way. There is no guarantee that our minds match the world here. 9 This is one of those cases (such as with the question of what is art, or what is pornography) where I am more confi dent about my judgment in particular cases than I am with any general formula that purports to cover every possible case. 10 This claim of exhaustivity will prove contentious. Some infl ationists (for instance, Armstrong 1983) would expand the realm of occurrent existence to include entities beyond history, and would go on to ground the laws in these additional occurrent entities. For further discussion see section 3.2. 11 It is worth distinguishing the reductionist’s general thesis, as I am explicating it, from the thesis that Lewis labels ‘Humean supervenience’ (1986b: ix–x). The reductionist thesis is both stronger and weaker than Humean supervenience. The reductionist thesis is stronger in that it is supposed to hold with metaphysical necessity, whereas Humean supervenience is only supposed to hold at a restricted region of logical space. But the reductionist thesis is weaker in that it makes no claims to locality or to reduction of whole to part, whereas Humean supervenience adds that the whole is grounded in its parts (thus the Humean would add that history itself reduces to the arrangement of the little point events). 12 Sider draws a similar moral: “What seems common to all the cheats is that irreducibly hypothetical [/modal] properties are postulated, whereas a proper ontology should invoke only categorical, or occurrent, properties and relations” (2001: 41; see also Armstrong 1997: 80–3). 13 Some of the more important attempts at a conceptual analysis of causation include Mackie (1974), Lewis (1986b), and Mellor (1995). For a discussion of some systematic counterex- amples (as well as a failed attempt at further analysis), see Schaffer (2001a). 14 Premise 2 is also explicitly invoked by the infl ationist Tooley: “If causal facts are logically supervenient upon non-causal facts, then it would seem that it must be in principle pos- sible to analyze causal concepts in non-causal terms” (1987: 177). Though no further argument is given. 15 This claim of conceptual centrality is explicit in Hume, who spoke of the concepts of causation, resemblance, and contiguity as “the only ties of our thoughts, . . . to us the cement of the universe . . .” (1978: 662) The centrality claim resurfaces in Carroll’s infl a- tionism: “With regard to our total conceptual apparatus, causation is the center of the center” (1994: 118; see also pp. 81–5). 16 This proposal is defended in Davidson (1969). 17 See Schaffer (2001b) for a defense of this individuation principle for tropes (particular properties), which may be identifi ed with events. The hardest case for this principle is the seeming possibility of multiple tropes/events with the same intrinsic natures piled in one place (Daly 1997: 154). I reply that the alleged piling either makes for an intensive difference or not. If not, I see no reason to believe that more than one trope/event is present. If so, I see no reason to believe that a pile of low-intensity tropes/events is present, rather than one high-intensity trope/event. In any case, the causal individuation principle surely does worse with respect to piling intuitions. Why can’t there be a world containing the following closed causal sequence: (i) e1 causes e2a and e2b, and (ii) e2a and e2b cause e3? Here e2a and e2b are causally piled – they have the exact same causes (e1) and effects (e3). Still they may be located in different places and have distinct natures: e2a might be a fl ash of green over here, while e2b might be a thunderous boom over there. This is excellent reason to think that more than one event is present, or so it seems to me. Causation and Laws of Nature: Reductionism 101 18 Thus Shoemaker has argued:
[I]f the properties and causal potentialities of a thing can vary independently of one another, then it is impossible for us to know (or have any good reason for believing) that something has retained a property over time, or that something has undergone a change with respect to the properties that underlie its causal powers. (1980: 215)
19 For further discussion, see Schaffer 2004 (esp. §3). There I consider the leading accounts of our knowledge of the external world, and conclude that “[knowledge of intrinsic natures] is possible in the same way that knowledge of the external world is possible, whatever that may be.” (p. 228) 20 Armstrong (1983: 133), Tooley (1987: 199–202), and Carroll (1994: 134–41) all provide examples of this sort. 21 So what does cause the prince to transform into a frog? Given the case as described, I would answer that both spells caused the prince to transform (both spells independently raised the chance of the transformation, after all). Though the case may be modifi ed to rule out the ‘both’ answer, by making it such that (i) when two spells both work the effect is enhanced (the prince would become extra-green with an extra-long tongue, say), and (ii) the enhanced effect does not obtain. In such a case I would answer that one of the spells caused the prince to transform, though it is ontologically indeterminate as to which. In some cases there is simply no fact of the matter. That is OK. Fundamental reality remains perfectly determinate. 22 The defender of (7) might offer the counterargument that (i) the three distinct scenarios are each conceivable, (ii) conceivability entails (or at least provides strong evidence for) possibility, so (iii) the three distinct scenarios are each possible (or at least there is strong evidence for such). To this I would reply that we must distinguish off-hand from ideal conceivability. Many things are off-hand conceivable that turn out to be impossible, such as trisecting an angle with ruler and compass. The only plausible link from conceivability to possibility is via ideal conceivability (Chalmers 2002). So I would grant that the three distinct scenarios are off-hand conceivable, but draw no conclusions from that. And I would deny that the three distinct scenarios are ideally conceivable – I am claiming that they are the result of conceptual error. 23 Something like this argument is present in Hume’s skeptical refl ection on the notion of necessary connection: “One event follows another; but we never can observe any tie between them. They seem conjoined but never connected” (1975:). Indeed, something like this argument seems to be the main impetus to reduction in the literature, both for causa- tion and for lawhood (see section 3.1). 24 Or at least, only the person who is skeptical of knowledge generally would deny (10). But such a skeptic should already be accustomed to postulating entities she claims no knowl- edge of, so she should not fi nd causal infl ationism any worse. 25 See Tooley (1987) for a sophisticated inferential account of causal knowledge. 26 See Fales (1990: esp. ch. 1) and Armstrong (1997: esp. 211–16) for a defense of this idea, though see Hume (1975: 352–9) for anticipatory criticism. 27 These examples are from Strawson (1985: 123). See Anscombe (1993: esp. 92–3) for further defense of this idea. 28 Point of clarifi cation: I am not opposing the use of epistemological arguments in meta- physics, but only the use of bad epistemological arguments. If (11) were true, then causal reductionism would follow by modus ponens. The occurrence of epistemic terminology does not invalidate modus ponens! 29 Or at least, I take it no one would deny the ontological economy aspect of (13). The theo- retical fathomability aspect might be more contested. But this will not matter for the argument of the main text. Jonathan Schaffer 102 30 In this vein, Russell dismissed causation as a relic of ‘Stone Age metaphysics’, since: “In the motions of mutually gravitating bodies, there is nothing that can be called a cause, and nothing that can be called an effect; there is merely a formula.” (1992: 202) See Quine (1966) for a similar claim. 31 See Reichenbach (1956: 147–9) and Salmon (1984: esp. 141–4) for the relevant arguments. The core idea is that there are some worldlines (which Reichenbach calls “unreal sequences” and which Salmon calls “pseudo processes”) that can move faster than light. Salmon gives the example of a rotating beacon in the center of a very large dome – if the beacon spins fast enough, and the wall of the dome is distant enough, then the spot of light moving around the wall can move at superluminal velocities. The reconciliation with special rela- tivity is that such worldlines are not capable of being used for signaling, which is a causally loaded notion. (It is worth noting here, in anticipation of the argument of the next para- graph, that both Reichenbach and Salmon advocate reductive accounts of causation.) 32 Note that the converse does not hold. If a certain notion proves needed in our best attempt at a systematic understanding of the world, this may be because what this notion concerns is ontologically basic, or it may be because this notion constitutes an irreplaceable con- ceptual shortcut for us. The latter represents the position I will take on laws (see section 4.3). 33 Point of clarifi cation: Russell argued for the elimination of causation, calling it “a relic of a bygone age, surviving, like the monarchy, only because it is erroneously supposed to do no harm” (1992: 193). I am not embracing Russell’s eliminativism (after all, I am defending reductionism here). Rather, I am arguing that if the infl ationist could establish a failure of reduction, then she would face Russell’s rejoinder that such an irreducible and scientifi cally irrelevant relation deserves to be eliminated. 34 An example of such an infl ationist is Maudlin (2004 and manuscript). When I mentioned reserving R3 as a fallback position (section 2), this was the sort of view I had in mind. 35 The conceptual centrality of lawhood is emphasized by Carroll: “Lawhood is conceptually intertwined with many other blatantly modal concepts that all have a massive role to play in our habitual ways of thinking and speaking” (1994: 3). 36 Tooley (1977: 669–72), van Fraassen (1989: 46–7), and Carroll (1990: 215–18; 1994: ch. 3.1) are among the many who have provided examples of this sort. 37 Earman (1984: 195) provides the following “empiricist loyalty test on laws”: (E1) For any w1 and w2, if w1 and w2 agree on all occurrent facts, then w1 and w2 agree on laws. Earman’s E1 is a consequence of reductionism and the further claim (section 2) that history exhausts occurrent existence. 38 It is contentious whether the empty world is a genuine possibility. But the “embarrassment of riches” problem arises for the infl ationist either way – the empty world is just the most dramatic case. 39 The intuition about lawhood seems even fl imsier than the analogous intuition about causa- tion that arises in the causal differences argument of (7)–(9). Lawhood is not quite as central a notion as causality. Indeed (as will be discussed shortly), our concept of lawhood is a relatively recent introduction by seventeenth century natural philosophers, involving the dubious idea of divine decrees for how the world should work. 40 See van Fraassen (1989: 5–7), and for a more extended historical discussion, Milton (1998). Here is what Milton concludes: “By the end of the sixteenth century the idea of God ordaining laws of nature had become suffi ciently familiar. . . . It was Descartes who more than anyone created the modern idea of a law of nature” (1998: 699) 41 Here I am following Beebee, who argues that the intuitive argument for 25 is simply ques- tion-begging: “[S]uch thought experiments do not succeed in fi nishing off the Humean conception of laws, because they presuppose a conception of laws which Humeans do not Causation and Laws of Nature: Reductionism 103 share: a conception according to which the laws govern what goes on in the universe” (2000: 573). Beebee goes on to trace the idea of governing to the theological conception of laws (pp. 580–1) 42 This argument is provided by Maudlin (manuscript, pp. 25–6). 43 There is the further worry that this aspect of scientifi c practice might have historical roots in the theological conception of laws, where the implicit idea is of modeling different acts of creation (this suspicion is voiced in Beebee [2000: 581]). Loewer (1996, pp. 116–17) provides the further suggestion that this portion of scientifi c practice may be explained away (without great loss to science) as a hasty generalization from acceptable ways of applying the actual laws to subsystems of the actual world. 44 Strictly speaking, the infl ationist should require that the electron and proton be embedded in a bigger world where enough is going on to ground the attribution of chance to the proton. But I take it that this should not raise any problems. Perhaps the easiest embed- ding strategy is to move to worlds with lots of electrons moving in straight lines, some random portion of which are accompanied by protons popping into existence and moving as they are wont. 45 This mode of argument is due to Carroll (1990: 215–18) who later develops it in detail (1994: ch. 3.1 and app. B). My principle LUC corresponds to Carroll’s SC*. Carroll mainly focuses on the principle he labels SC (If it is nomologically possible that p and nomologi- cally necessary that [if p then q], then had p been the case then q would have been the case), claiming that SC*(=LUC) is a consequence of SC suffi cient to trouble the reductionist. But the derivation of SC* from SC requires the following inference: “Since Q is a law, Q is a law in every possible world with the same laws as the actual world, and so it is physi- cally necessary that Q is a law” (Carroll 1994: 59). This is an inference from ᮀq (q is a law) to ᮀᮀq (it is physically necessary that q is a law). That inference is only valid in modal systems that include the K4 axiom ᮀp → ᮀᮀp. See the next footnote for further discussion. 46 K4 is a normal modal system (meaning that it includes as theorems all the tautologies and distribution axioms, and is closed under modus ponens, substitution, and necessitation) augmented with the axiom ᮀp → ᮀᮀp. This axiom functions to make accessibility transi- tive, which is what validates LUC. One can generate countermodels to LUC in normal modal systems without the transitivity-generating K4 axiom. For instance, here is a coun- termodel in T (a normal modal system augmented with the refl exivity-generating axiom ᮀp → p, which I think is the best modal logic for modeling lawhood). Set the model
References
Anscombe, G. E. M. 1993. “Causality and Determination,” in E. Sosa and M. Tooley, eds., Causation (Oxford: Oxford University Press), pp. 88–104. ——. 1983. What is a Law of Nature? (Cambridge: Cambridge University Press). ——. 1997. A World of States of Affairs (Cambridge: Cambridge University Press). Beebee, H. 2000. “The Non-Governing Conception of Laws of Nature,” Philosophy and Phenomenological Research 81: 571–94. Carroll, J. 1990. “The Humean Tradition,” Philosophical Review 99: 185–219. ——. 1994. Laws of Nature (Cambridge: Cambridge University Press). Chalmers, D. 2002. “Does Conceivability Entail Possibility?” in T. Szabó Gendler and J. Hawthorne, eds., Conceivability and Possibility (Oxford: Oxford University Press), pp. 145–200. Causation and Laws of Nature: Reductionism 105 Daly, Chris. 1997. “Tropes,” in D. H. Mellor and A. Oliver, eds., Properties (Oxford: Oxford University Press), pp. 125–39. Davidson, D. 1969. “The Individuation of Events,” in Essays on Actions and Events (Oxford: Oxford University Press), pp. 163–80. Dretske, F. 1977. “Laws of Nature,” Philosophy of Science 44: 248–68. Earman, J. 1984. “Laws of Nature: The Empiricist Challenge,” in R. J. Bogdan, ed., D. M. Armstrong (Dordrecht: Reidel Publishing), pp. 191–223. Earman, J. and Roberts, J. 2005. “Contact with the Nomic: A Challenge for Deniers of Humean Supervenience about Laws of Nature. Part II: The Epistemological Argument for Humean Supervenience,” Philosophy and Phenomenological Research 71: 253–86. Fales, E. 1990. Causation and Universals (London: Routledge & Kegan Paul). Fine, K. 1994. “Ontological Dependence,” Proceedings of the Aristotelian Society 95: 269–90. ——. 2001. “The Question of Realism,” Philosophers Imprint 1: 1–30. Hume, D. 1975 An Enquiry Concerning Human Understanding (Oxford: Oxford University Press). ——. 1978. A Treatise of Human Nature (Oxford: Oxford University Press). Lewis, D. 1973. Counterfactuals (Cambridge, MA: Harvard University Press). ——. 1983. “New Work for a Theory of Universals,” Australasian Journal of Philosophy 61: 343–77. ——. 1986a. On the Plurality of Worlds (Oxford: Basil Blackwell). ——. 1986b. Philosophical Papers, vol. 2 (Oxford: Basil Blackwell). ——. 1994. “Humean Supervenience Debugged,” Mind 103: 473–490. Loewer, B. 1996. “Humean Supervenience,” Philosophical Topics 24: 101–27. ——. 2001. “From Physics to Physicalism,” in C. Gillet and B. Loewer, eds., Physicalism and its Discontents (Cambridge: Cambridge University Press), pp. 37–56. Lowe, E. J. 2005. “Ontological Dependence.” Stanford Encyclopedia of Philosophy.
Causation and Laws of Nature: Reductionism 107
CHAPTER THREE MODALITY AND POSSIBLE WORLDS
3.1 “Concrete Possible Worlds,” Phillip Bricker 3.2 “Ersatz Possible Worlds,” Joseph Melia
The twentieth-century writer Rex Stout wrote detective fi ction, but he might have become a real detective instead. In some other possible world, he really does become a detective. In yet another world, Stout has yet another occupation: he is a salesman. For every occupation that Stout could have had, there is a possible world in which Stout has that occupation. Many things vary between different possible worlds: Stout has different occupations, different clothes, different hair color, different friends, and so on. The only things that hold constant in all possible worlds are the necessary truths: in every possible world, Stout is either a salesman or he isn’t. Philosophers have found it convenient to speak in this way of “possible worlds,” but what are possible worlds, really? Phillip Bricker argues that we should take possible worlds talk at face value. Other possible worlds, containing other Rex Stouts with their different occupations, clothes, and friends, really exist. Joseph Melia disagrees; we should instead regard talk of possible worlds as really being talk of more mundane entities; for example, stories that describe the alternate occupations of Rex Stout and other non-actual matters.
CHAPTER 3.1
Concrete Possible Worlds
Phillip Bricker
1 Introduction
Open a book or article of contemporary analytic philosophy, and you are likely to fi nd talk of possible worlds therein. This applies not only to analytic metaphysics, but also to areas as diverse as philosophy of language, philosophy of science, epistemol- ogy, and ethics. Philosophers agree, for the most part, that possible worlds talk is extremely useful for explicating concepts and formulating theories. They disagree, however, over its proper interpretation. In this chapter, I discuss the view, championed by David Lewis, that philosophers’ talk of possible worlds is the literal truth.1 There exists a plurality of worlds. One of these is our world, the actual world, the physical universe that contains us and all our surroundings. The others are merely possible worlds containing merely possible beings, such as fl ying pigs and talking donkeys. But the other worlds are no less real or concrete for being merely possible. Fantastic? Yes! What could motivate a philosopher to believe such a tale? I start, as is customary, with modality.2 Truths about the world divide into two sorts: categorical and modal. Categorical truths describe how things are, what is actu- ally the case. Modal truths describe how things could or must be, what is possibly or necessarily so. Consider, for example, the table at which I am writing. The table has numerous categorical properties: its color, perhaps, and its material composition. To say that the table is brown or that it is made of wood is to express a categorical truth about the world. The table also has numerous modal properties. The table could have been red (had it, for example, been painted red at the factory), but it could not, it seems, have been made of glass, not this very table; it is essentially made of wood. Just where to draw the line between the categorical and the modal is often disputed. But surely (I say) there is some level – perhaps fundamental physics – at which the world can be described categorically, with no admixture of modality. Now, sup- pose one knew the actual truth or falsity of every categorical statement. One might nonetheless not know which truths are necessary or which falsehoods are possible. One might be lacking, that is, in modal knowledge. In some sense, then, the modal transcends the categorical. And that’s trouble; for modal statements are problematic in a way that categorical statements are not. I see that the table is brown, but I do not see that it is possibly red. I can do empirical tests to determine that it is made of wood, but no empirical test tells me that it is essentially made of wood. By observation, I discover only the categorical properties of objects, not their modal properties. That makes special trouble for the empiricist, who holds that all knowledge of the world must be based on observation. But empiricist or not, modal properties are mysterious: they do not seem to be among the basic or fundamental ingredients that make up our world. What to do? Should we turn eliminativist about modality, holding that modal statements are unintelligible, or, at any rate, that their communicative purpose is not descriptive?3 That would be implausible. We assign truth and falsity to modal statements in principled ways; we reason with modal statements according to their own peculiar logic. No, we must hold that modal statements are descriptively meaningful, but not fundamental. Thus begins the search for an analysis of modal statements, the attempt to provide illuminating truth conditions for modal statements without just invoking more modality. Consider this. Modal statements can be naturally paraphrased in terms of possible worlds. For example, instead of saying “It is possible that there be blue swans,” say “In some possible world there are blue swans.” Instead of saying “It is necessary that all swans be birds,” say “In every possible world all swans are birds.” When para- phrased in this way, the modal operators ‘it is possible that’ and ‘it is necessary that’ become quantifi ers over possible worlds. Intuitively, these paraphrases are merely a façon de parler, and not to be taken with ontological seriousness. But perhaps the ease with which we can produce and understand these possible worlds paraphrases suggests something different: the paraphrases provide the sought-after analyses of the modal statements; they are what our modal talk has been (implicitly) about all along.4 And if that is so, then we have a reduction of the modal to the categorical after all: although the modal properties of this world transcend the categorical properties of this world, they are determined by the categorical properties of this world, and other possible worlds. But the introduction of possible worlds may seem to raise more ques- tions than it answers. What are these so-called worlds? What is their nature, and how are they related to our world? Philosophers who believe in possible worlds divide over whether worlds are abstract or concrete. I use the terms ‘abstract’ and ‘concrete’ advisedly: there are (at least) four different ways of characterizing the abstract/concrete distinction, making the terms ‘abstract’ and ‘concrete’ (at least) four ways ambiguous.5 Fortunately, however, on the Lewisian approach to modality I am considering, the worlds turn out (with some minor qualifi cations) to be “concrete” on all four ways of drawing the distinction.
(C1) Worlds (typically) have parts that are paradigmatically concrete, such as donkeys, and protons, and stars. (C2) Worlds are particulars, not universals; they are individuals, not sets. (C3) Worlds (typically) have parts that stand in spatiotemporal and causal rela- tions to one another. Phillip Bricker 112 (C4) Worlds are fully determinate;6 they are not abstractions from anything else.
It is convenient, then, and harmless, to call worlds “concrete” if they satisfy all four conditions listed above. The concrete worlds taken altogether I call the modal realm, or, following custom, logical space.7 The denizens of logical space – the worlds and their concrete parts – are called possibilia. In one sense, concrete possible worlds are like big planets within the actual world: two concrete worlds do not have any (concrete) objects in common; they do not overlap. Thus, you do not literally exist both in the actual world and in other merely possible worlds, any more than you literally exist both on Earth and on other planets in our galaxy. Instead, you have counterparts in other possible worlds, people qualita- tively similar to you, and who play a role in their world similar to the role you play in the actual world. This constrains the analysis of modality de re – statements ascribing modal properties to objects. When we ask, for example, whether you could have been a plumber, we are not asking whether there is a possible world in which you are a plumber – that is trivially false (supposing you are not in fact a plumber), since you inhabit only the actual world. Rather, we are asking whether there is a possible world in which a counterpart of you is a plumber.8 (More on this in section 5 below.) In another sense, however, concrete worlds are quite unlike big planets within the actual world. Each possible world is spatiotemporally and causally isolated from every other world: one cannot travel between possible worlds in a spaceship; one cannot view one world from another with a telescope. But although this makes other possible worlds empirically inaccessible to us in the actual world, it does not make them cognitively inaccessible: we access other worlds through our linguistic and mental representations of ways things might have been, through the descriptions we formu- late that the worlds satisfy. I say that merely possible concrete worlds, no less than planets within the actual world, exist. I do not thereby attribute any special ontological status to the worlds (or planets): whatever has any sort of being “exists,” as I use the term; ‘existence’ is coextensive with ‘being’. Of course, merely possible concrete worlds do not actually exist. For the realist about concrete worlds, existence and actual existence do not coincide. In most ordinary contexts, no doubt, the term ‘exists’ is implicitly restricted to actual things – as, for that matter, is the term ‘is’. The phenomenon of implicit contextual restriction allows us to truly say, in an ordinary context, “fl ying pigs do not exist” or “there are no fl ying pigs,” without thereby denying the existence of concrete non-actual worlds teeming with fl ying pigs. It is the same phenomenon that allows us to truly say, when there is no beer in our fridge, “there is no beer,” without denying that there are other fridges in the world packed with beer. Ordinary assertions of non-existence, then, do not count against realism about concrete worlds. Call approaches to modality that analyze modality in terms of concrete possible worlds and their parts Lewisian approaches. I take the following four theses to be characteristic of Lewisian approaches to modality:
1 There is no primitive modality. 2 There exists a plurality of concrete possible worlds. Concrete Possible Worlds 113 3 Actuality is an indexical concept.9 4 Modality de re is to be analyzed in terms of counterparts, not transworld identity.
In what follows, I devote one section to each of these theses. I write as an advocate for Lewisian approaches, and feel under no obligation to give opposing views equal time. For each thesis, I take Lewis’s interpretation and defense as my starting point. I then consider and endorse alternative ways of accepting the thesis, some of which disagree substantially with Lewis’s interpretation or defense. There is more than one way to be a Lewisian about modality.
2 No Primitive Modality
The rejection of primitive modality is a central tenet of Lewisian approaches. It moti- vates the introduction of possible worlds, the most promising avenue of analysis. And it motivates taking possible worlds to be concrete: Lewis’s most persistent complaint against accounts of worlds as abstract is that they must invoke primitive modality in one form or another. But for all the talk of rejecting primitive modality by Lewis and others, there is no clear agreement as to just what this means. Indeed, I think three different and independent theses have been taken to fall under the “no primitive modality” banner. Although Lewis accepts all three theses, only one of them is truly central to the Lewisian approach. Before turning to discuss that central thesis, I will briefl y mention the other two. First off, it is natural to interpret ‘no primitive modality’ as a supervenience thesis: the modal supervenes on the categorical. To say that modal statements supervene on categorical statements is to say that there can be no difference as to how things are modally without some difference as to how they are categorically. Or, cashed out in terms of possible worlds: whenever two possible worlds differ in their modal features, they differ in their categorical features as well. Thus, for example, no two possible worlds differ just with respect to brute dispositional properties, or primitive causal relations. This supervenience thesis is central to Lewis’s (broadly) Humean analyses of laws, causation, and counterfactuals, of the physical and causal modalities (see chapter 2.2). But the thesis is not directed at the metaphysical (or logical) modality that is the target of our current analytic endeavor. I therefore set it aside. A second way one might understand the ‘no primitive modality’ thesis is as what Lewis calls a principle of recombination. His initial formulation of the principle is: “[A]nything can coexist with anything else, and anything can fail to coexist with anything else” (1986: 88). The fi rst half, when spelled out, says that any two (or more) things, possibly from different worlds, can be patched together in a single world. To illustrate: if there could be a unicorn, and there could be a dragon, then there could be a unicorn and a dragon side by side. Since worlds do not overlap, a unicorn from one world and a dragon from another cannot themselves exist side by side. So the principle is to be interpreted in terms of intrinsic duplicates: at some world, a duplicate of the unicorn and a duplicate of the dragon exist side by side. The second half of the principle of recombination, spelled out in terms of worlds and duplicates, says Phillip Bricker 114 this: whenever two (non-overlapping) things coexist at a world, neither of which is a duplicate of a part of the other, there is another world at which a duplicate of one exists without a duplicate of the other. To illustrate: since a talking head exists con- tiguous to a living human body, there could exist an unattached talking head, separate from any living body. More precisely: there is a world at which a duplicate of the talking head exists but at which no duplicate of the rest of the living body exists. According to Lewis, the principle of recombination expresses “the Humean denial of necessary connections between distinct existents” (Lewis 1986: 87). But two caveats are needed. First, only the second half, strictly speaking, embodies a denial of neces- sary connections; the fi rst half embodies instead a denial of necessary exclusions. And, second, the thesis that the modal supervenes on the categorical is an alternative way to capture the denial of necessary connections, one that may be closer to Hume’s intent; it denies that there are necessary connections in the world, ontological inter- lopers somehow existing over and above the mere succession of events, and somehow serving as ground for modal truths about powers or laws or causation. Lewis sometimes charges those who reject principles of recombination with being committed to primitive modality.10 But I do not think that a principle of recombina- tion is the right way to capture the ‘no primitive modality’ thesis. Violations of recombination impose a modal structure on logical space, allowing that the existence of some possibilia necessarily entail or exclude the existence of other possibilia; but the structure imposed need not be primitive modal structure. It may be that the viola- tions can all be accounted for in non-modal terms, that necessary connections and exclusions only occur when some non-modal condition is satisfi ed. Thus, contra what Lewis suggests, violations of recombination are not – or at least not by themselves – primitive modality. The fi rst two interpretations of the ‘no primitive modality’ thesis each put con- straints on logical space, though in opposite ways. The supervenience thesis demands that there not be too many worlds, that there never be two worlds that differ modally without differing non-modally. Principles of recombination demand that there not be too few worlds, that there always be enough worlds to represent all the different ways of recombining. The third interpretation – the one we’ re after – is different: it puts constraints on our theorizing about logical space, not (directly) on logical space itself. It demands that our total theory, our best account of the whole of reality, can be stated without recourse to modal notions, that the (primitive) ideology of our total theory be non-modal.11 Can the Lewisian meet this demand? First, we need to know what terms of our total theory count as non-modal. Here I will suppose this includes the Boolean con- nectives (‘and’, ‘or’, ‘not’) and unrestricted quantifi ers (‘every’, ‘some’) of logic, the ‘is a part of’ relation of mereology, the ‘is a member of’ relation of set-theory, the ‘is an instance of’ relation of property theory, and a (second-order) predicate applying to just those properties and relations that are fundamental, or perfectly natural.12 I will also suppose that the notion of a spatiotemporal relation is non-modal; perhaps it can be given a structural analysis in terms of the above.13 Now, let us suppose that the Lewisian has succeeded in providing possible worlds paraphrases for the vast panoply of modal locutions. That still leaves the notion of a possible world itself, an ostensibly modal notion. It won’t do simply to take this notion as primitive. At best, Concrete Possible Worlds 115 that would reduce the number of modal notions to one. If primitive modality is to be eliminated, the Lewisian must provide an analysis of ‘possible world’ in non- modal terms. It might appear that analyzing ‘possible world’ involves two separate tasks: fi rst, analyzing the notion of world; then, distinguishing those worlds that are possible from those that are not. This second task, however, would appear to land the Lewisian in vicious circularity: possibility is to be analyzed in terms of possible worlds, which in turn is to be analyzed in terms of possibility, and round and round and round.14 But the threat of circularity is bogus because there is no second task to perform. On a conception of possible worlds as concrete, there are no impossible worlds. For suppose there were a concrete world at which both p and not-p, for some proposition p. Then there would be a property corresponding to p such that the world both had and didn’t have the property. Contradictions could not be confi ned to impossible worlds; they would infect what is true simpliciter, thereby making our total theory contradictory.15 The law of non-contradiction, then, demands that the Lewisian reject impossible worlds. But if there are no impossible worlds, the ‘possible’ in ‘possible world’ is redundant, and no separate analysis is needed to pick out the worlds that are possible from the rest. Let us, then, focus on the one and only task: analyzing the notion of world. To accomplish this task, it suffi ces to provide necessary and suffi cient conditions for when two individuals are worldmates, are part of one and the same world. Lewis’s proposal is this: individuals are worldmates if and only if they are spatiotemporally related to one another, that is, if and only if every part of one stands in some distance (or interval) relation – spatial, temporal, spatiotemporal – to every part of the other (Lewis 1986: 71). This leads immediately to the following analysis of the notion of world: a world is any maximal spatiotemporally interrelated individual – an individual all of whose parts are spatiotemporally related to one another, and that is not included in a larger individual all of whose parts are spatiotemporally related to one another. On this account, a world is unifi ed by the spatiotemporal relations among its parts. If one further assumes with Lewis that being spatiotemporally related is an equivalence relation (refl exive, symmetric, and transitive), it follows that each individual belongs to exactly one world: the sum (or aggregate) of all those individuals that are spatio- temporally related to it. Now, the point to emphasize for present purposes is this: if Lewis’s analysis is accepted, the notion of world has been characterized in non-modal terms, and the claim to eschew primitive modality has been vindicated. The acceptability of Lewis’s analysis of world hinges on the acceptability of his analysis of the worldmate relation. One direction of the analysis (suffi ciency) is uncontroversial. Whatever stands at some spatiotemporal distance from us is part of our world; or, contrapositively, non-actual individuals stand at no spatiotemporal distance from us, or from anything actual. In general: every world is spatiotemporally isolated from every other world. According to the other direction of the analysis (necessity), worlds are unifi ed only by spatiotemporal relations; every part of a world is spatiotemporally related to every other part of that world. This direction is more problematic, for at least three reasons. Although I believe Lewis’s account needs to be modifi ed to solve these problems, in each case the modifi cation I would suggest does not require introducing primitive modality. Phillip Bricker 116 First, couldn’t there be worlds that are unifi ed by relations that are not spatiotem- poral? Indeed, it is controversial, even with respect to the actual world, whether entities in the quantum domain stand in anything like spatiotemporal relations to one another; the classic account of space-time may simply break down. I have defended elsewhere a solution that Lewis considered but (tentatively) rejected: individuals are worldmates if and only if they are externally related to one another, that is, if and only if there is a chain of perfectly natural relations (of any sort) extending from any part of one to any part of the other (Bricker 1996). This analysis quantifi es over all perfectly natural relations rather than just the spatiotemporal relations, but it is none the worse for that with respect to primitive modality. Second, isn’t it possible for there to be disconnected space-times, so-called “island universes”? Couldn’t there be a part of actuality spatiotemporally and causally isolated from the part we inhabit? Lewis must answer “no.” When Lewis’s analysis of world is combined with the standard analysis of possibility as truth at some world, island universes turn out to be impossible: at no world are there two disconnected space- times. But although Lewis rejects the possibility of island universes, he is uneasy, and for good reason: principles of recombination that Lewis seems to accept entail that island universes are possible after all.16 Of course, if one took the worldmate relation to be primitive, one could simply posit as part of the theory that some worlds are composed of disconnected space-times. But that’s no good. A primitive worldmate relation is primitive modality: what is possible – for example, the possibility of island universes – depends on how the worldmate relation is laid out in logical space. Is there a way to allow for the possibility of island universes without invoking primitive modality? This time, I think, the best strategy is to amend the analysis of modality instead of the analysis of world. The modal operators should be taken to quantify over worlds and pluralities of worlds. For example, to be possible is to be true at some world, or some plurality of worlds. What is true at a plurality of worlds is, intuitively, what would be true if all the worlds in the plurality were actualized. If a plurality of two or more worlds were actualized, then actuality would include two or more discon- nected parts; and so, on the amended analysis, island universes turn out to be possible.17 A third problem affl icts Lewisians who are also Platonists of a certain kind, namely, those who believe that in addition to the modal realm consisting of the isolated con- crete worlds there is a mathematical realm consisting of isolated mathematical systems of abstract entities. Most Platonists (although not Lewis) believe in at least one such system: the pure sets, externally related to one another in virtue of the structure imposed by the membership relation. But I, for one, believe also in sui generis numbers, externally related to one another in virtue of the structure imposed by the successor relation, in sui generis Euclidean space, and much, much more. But, then, on the (amended) analysis of world being considered, the pure sets comprise a world, the natural numbers comprise a world, and so on for all the mathematical systems one believes in. That makes metaphysical possibility, which is analyzed as a quantifi er over worlds, depend on what mathematical systems there are, and that seems wrong.18 The Lewisian, then, needs to fi nd a way to demarcate modal reality from mathematical reality, a way that that does not invoke primitive modality. Concrete Possible Worlds 117 It might seem that the conditions (C1)–(C4) used to characterize what makes a world concrete could do the job. For the mathematical systems, one might hold, are abstract in virtue of violating (C4): they are not fully determinate; it is neither true nor false of numbers, for example, that they are red, or weigh ten grams. But that is not how I see it. Indeed, the pure sets, or sui generis numbers, or Euclidean points have no intrinsic qualitative nature, but not because they are somehow abstractions from something else.19 Rather, they determinately fail to instantiate every qualitative property.20 Their nature is purely relational, but it is no less determinate for that. But perhaps there is a simple fi x. I think there is a fi fth way in which entities can be said to be concrete: concrete entities have an intrinsic qualitative nature in virtue of instantiating, or having parts that instantiate, perfectly natural properties. This pro- vides a fi fth condition to be satisfi ed by the concrete worlds:
(C5) A world is a sum of individuals, each of which instantiates at least one per- fectly natural property.21
Incorporating (C5) into the analysis of world provides the needed distinction between the modal and the mathematical realm. And since the notion of a perfectly natural property is non-modal, we have not had to invoke primitive modality to do the job. The elimination of primitive modality is a central goal of Lewisian approaches. To achieve this goal, the Lewisian must make a case for the following controversial claims. First, there is the claim that there exists a plurality of concrete worlds; that claim will be the focus of section 3. Second, there is the claim that the analysis of modality in terms of concrete worlds is materially and conceptually adequate; that claim will be put to the test in section 4 with respect to the analysis of actuality, and in section 5 with respect to the analysis of modality de re. Finally, there is the claim, noted above, that there are no impossible concrete worlds. That claim rests ultimately on a defense of classical logic, a topic too far afi eld to pursue further here.
3 Concrete Worlds Exist
Why believe in concrete worlds other than the actual world? Throughout his career, Lewis held to a broadly Quinean methodology for deciding questions of existence. Roughly, we are committed to the existence of those entities that are quantifi ed over by the statements we take to be true.22 And we should take those statements to be true that belong to the best total theory, where being “best” is in part a matter of being fruitful, simple, unifi ed, economical, and of serving the needs of common sense, science, and systematic philosophy itself. What we should take to exist, then, is deter- mined by criteria both holistic and pragmatic. Early on, Lewis applied the Quinean methodology directly to statements we accept in ordinary language (Lewis 1973: 84). We accept, for example, not only that “things might be otherwise than they are,” but also that “there are many ways things could have been besides the way they actually are.” This latter statement quantifi es explicitly over entities called “ways things could have been,” entities that Lewis identifi es with Phillip Bricker 118 concrete possible worlds. But that identifi cation is far from innocent. It was soon pointed out (in Stalnaker (1976) ) that the phrase ‘ways things could have been’ seems to refer, if at all, to abstract entities – perhaps uninstantiated properties – not to concrete worlds. Indeed, it is doubtful that any statements we ordinarily accept quan- tify explicitly over concrete worlds. In On the Plurality of Worlds, Lewis abandoned any attempt to apply the Quinean methodology directly to ordinary language, and applied it instead to systematic phi- losophy. Concrete worlds, if accepted, improve the unity and economy of philosophical theories by reducing the number of notions that must be taken as primitive. Moreover, concrete worlds provide, according to Lewis, a “paradise for philosophers” analogous to the way that sets have been said to provide a paradise for mathematicians (because, given the realm of sets, one has the wherewithal to provide true and adequate inter- pretations for all mathematical theories). Here Lewis has in mind not just the use of possible worlds to analyze modality, but also their use in constructing entities to play various theoretical roles, for example, the meanings of words and sentences in seman- tics and the contents of thought in cognitive psychology.23 So, when asked “why believe in a plurality of worlds?” Lewis responds: “because the hypothesis is service- able, and that is a reason to think that it is true” (Lewis 1986: 3). Lewis does not claim, of course, that usefulness by itself is a decisive reason to believe: there may be hidden costs to accepting concrete worlds; there may be alter- natives to concrete worlds that provide the same benefi ts without the costs. Lewis’s defense of realism about concrete worlds, therefore, involves an extensive cost-benefi t analysis. His conclusion is that, on balance, his realism defeats its rivals: rival theories that can provide the same benefi ts all have more serious costs. I will not attempt here to summarize Lewis’s lengthy and intricate discussion.24 But I will say something about the general idea that belief in concrete worlds can be given a pragmatic foun- dation, and I will ask whether an alternative foundation is feasible. Lewis’s argument for belief in concrete worlds depends on the assumption that we should believe pragmatically virtuous theories, theories that, on balance, are more fruitful, simple, unifi ed, or economical than their rivals. Although this assumption is orthodoxy among contemporary analytic philosophers, I fi nd it no less troubling for that. It is one thing for a theory to be pragmatically virtuous, to meet certain of our needs and desires; it seems quite another thing for the theory to be true. On what grounds are the pragmatic virtues taken to be a mark of the true? It is easy to see why we would desire our theories to be pragmatically virtuous: the virtues make for theories that are useful, productive, easy to comprehend and apply. But why think that reality conforms to our desire for simplicity, unity, and the other pragmatic virtues? Believing a theory true because it is pragmatically virtuous leads to parochi- alism, and seems scarcely more justifi ed than, say, believing Ptolemaic astronomy true because it conforms to our desire to be located at the center of the universe. Such wishful thinking is no more rational in metaphysics than in science or everyday life. But if we reject a pragmatic foundation for belief in concrete worlds, what is there to put in its place? My hope is that there are general metaphysical principles that support the existence of concrete worlds, and that we can just see, on refl ection, that these principles are true. This “seeing” is done not with our eyes, of course, but with our mind, with a Cartesian faculty of rational insight. This faculty is fallible, to be Concrete Possible Worlds 119 sure, as are all human faculties. (Contra Descartes, I do not take the faculty to be invested with the imprimatur of an almighty deity.) But fallible or not, some such faculty is needed lest our claim to have a priori knowledge be bankrupt. Now, what general principles could play a foundational role for belief in a plurality of concrete worlds? I will consider, briefl y, two lines of argument. One way to argue for controversial ontology is to invoke a truthmaker principle: for every (positive) truth, there exists something that makes it true, some entity whose existence entails that truth.25 Truths don’t fl oat free above the ontological fray. They must be grounded in some portion of reality. For example, that Fido is a dog, if true, has Fido himself as a truthmaker. (Assuming, as is customary, that an animal belongs to its species essentially.) That some animals are dogs has multiple truthmakers: each and every animal that is a dog. (On the other hand, a negative truth, such as that no dog is a bird, is made true by the lack of false-makers, by the non-existence of any dog that is a bird.) A more controversial case: that there are infi nitely many prime numbers, I claim, is made true by the existence of the system of natural numbers. Mathematical truths have mathematical entities as truthmakers. Consider now a (positive) modal truth such as it is possible that there be unicorns. What could be a truthmaker for this truth? Not any actual unicorn: there aren’t any. Not actual ideas of unicorns, or other actual mental entities; for the possibility of unicorns doesn’t depend on whether any mind has ever conceived of unicorns, or even whether any mind has ever existed. What then? To fi nd the truthmakers for a statement, it helps to ask what the statement is about. That unicorns are possible appears to be about unicorns, if about anything at all. And, by the truthmaker prin- ciple, it is about something. But since there are no actual unicorns, that leaves only merely possible unicorns for it to be about: it has each and every possible unicorn as a truthmaker. Thus baldly presented, the truthmaker argument for concrete possibilia may fail to convince. Indeed, the argument would need to be supplemented in at least two ways. First, even if one grants that modal truths require possibilia for truthmakers, why hold that the possibilia in question must be concrete? Perhaps abstract possibilia can meet the demand for truthmakers. Filling this gap in the argument, it seems, must wait on a decisive critique of all abstract accounts of possibilia – a tall order. Second, the truthmaker principle is often restricted to contingent truths, and for good reason. A truthmaker for a statement is an entity whose existence entails that statement. As entailment is ordinarily understood in terms of possible worlds, one statement entails a second just in case every world at which the fi rst is true is a world at which the second is true. Thus understood, any statement entails a necessary truth, and so truthmaking for necessary truths becomes a trivial affair, devoid of ontological con- sequence. But the thesis that concrete worlds exist (with “exists” unrestricted) is a necessary truth. If the truthmaker principle is to apply to this thesis, truthmaking must be based on a more discriminating notion of entailment. It won’t do to take this dis- criminating notion of entailment as primitive, lest the “no primitive modality” thesis be violated. So, some non-standard account of truthmaking in terms of worlds will need to be developed – no easy task. Given these diffi culties with the truthmaker argument for concrete possibilia, I fi nd a different line of argument more promising, one that focuses on the nature of Phillip Bricker 120 intentionality. Intentionality, in the relevant philosophical sense, refers to a feature of certain mental states such as belief and desire: these states are always “directed” toward some object or objects; one doesn’t just believe or desire, one always believes or desires something.26 That some of our mental states have this feature is something we know a priori by introspection. We know, that is, a general principle to the effect that mental states with this feature – “intentional states” – are genuinely relational. The second line of argument, then, is that this general principle can serve as founda- tion for belief in concrete possible objects and possible worlds. Concrete possibilia are needed to provide the objects of our intentional states, to provide an ontological framework for the content of our thought. To illustrate how the argument might go, consider the intentional state of thinking about some object or objects. Suppose, for example, that I am now thinking about a dodecahedron made of solid gold. I can do this, of course, whether or not any such object actually exists. If there is, unbeknownst to me, an actual gold dodecahedron, then I am related to it in virtue of being in my current intentional state; it is an object of my thought. But what if there is no actual gold dodecahedron? Does that somehow prevent me from thinking about one? Of course not. In either case, I claim, thinking about is relational, and relations require relata. In the latter case, only merely possible gold dodecahedrons are available to be objects of my thought; I am related to possible but non-actual objects. However, if the relationality of intentional states is to serve as a foundation for a Lewisian account of worlds, at least three further claims require support. First, the objects of thought, even when merely possible, must instantiate the same qualitative properties as actual objects of thought. Suppose again that a merely possible gold dodecahedron is an object of my thought. Does this object instantiate the property of being gold? Or is it an abstract object that somehow represents the property of being gold? I say the former. It is one thing to think about a gold dodecahedron, another thing to think about some abstract simulacrum thereof. If I am thinking about a gold dodecahedron and thinking about is genuinely relational, then there is a gold dodeca- hedron that I am thinking about. That it is made of gold and shaped like a dodeca- hedron is independent of whether it is actual or merely possible. Indeed, nothing prevents actual and merely possible objects from being perfect qualitative duplicates of one another. But, second, more is needed if the objects of my thought are to count as concrete: they must not only instantiate qualitative properties, they must be fully determinate in all qualitative respects. How can that be? Intentional states such as thinking about do not seem to be determinate with respect to their objects. In thinking about a gold dodecahedron, I wasn’t thinking about a gold dodecahedron of any particular size. Should I say, then, that I was related by my thought to an object that has no defi nite size? No. It is one thing to think indeterminately about a gold dodecahedron, another thing to think about an indeterminate object. The indeterminacy is in the thinking, not the object of thought. I am related by my thought to a multitude of possible gold dodecahedrons with a multitude of different, but fully determinate, sizes.27 But still more is needed if the argument is to support a Lewisian account: each concrete object of thought must be part of a fully determinate concrete world. In thinking about a gold dodecahedron, I wasn’t thinking about how it is situated with Concrete Possible Worlds 121 respect to other objects. But that is just another aspect of the indeterminacy of my thought. Each possible gold dodecahedron has a determinate extrinsic nature; my thought doesn’t discriminate between differently situated gold dodecahedra, and it therefore ranges indeterminately over them all. Perhaps, as I believe, there exists in logical space a solitary gold dodecahedron that is a world all by itself. But then distinguish: it is one thing to think about a solitary gold dodecahedron, another thing to think about a gold dodecahedron without considering how it relates to other objects. In the former case, what I am thinking about stands in no spatial or temporal rela- tions to other objects; in the latter case, it is indeterminate whether what I am thinking about stands in spatial or temporal relations to other objects. In either case, the pos- sible gold dodecahedra that are objects of my thought belong to fully determinate concrete worlds. That, in barest outline, is how the relationality of thought could serve as founda- tion for a Lewisian account of concrete worlds. A thoroughgoing Quinean pragmatist would say, of course, that the thesis of the relationality of thought – or the truthmaker principle of the preceding argument, or any fundamental metaphysical principle – can support belief in concrete worlds only to the extent that its acceptance confers benefi ts on our total theory. But if I am right that the pragmatic virtues are never, in and of themselves, a mark of the true, then a “pragmatic foundation” is not to be had; indeed, it is a contradiction in terms. Contra what Lewis claims, that a belief is “serviceable” for the project of systematic philosophy provides no reason at all to hold it. Founding belief in concrete worlds instead on a (fallible) faculty of rational insight into matters metaphysical is controversial, to be sure, and in need of much development. But better a shaky foundation, I say, than no foundation at all.
4 Actuality is Indexical
Thus far, I have said very little about the notion of actuality. But some of the com- mitments of a Lewisian account are already clear. Since the Lewisian believes that merely possible worlds exist, she rejects the thesis that whatever exists is actual; that is to say, the Lewisian is a possibilist, not an actualist. Moreover, since the Lewisian believes that merely possible concrete worlds exist, she rejects any identifi cation of the actual with the concrete. Furthermore, since the Lewisian holds that actual things have qualitative duplicates in merely possible worlds, actuality cannot itself be any sort of qualitative property. What, then, is it? In virtue of what do actual things differ from their merely possible counterparts? The Lewisian needs a positive account of actuality. Lewis responds by proposing a defl ationary account of actuality. The actual world and the merely possible worlds are ontologically all on a par; there is no fundamental, absolute property that actual things have and merely possible things lack. Nonetheless, I speak truly when I call my world and my worldmates “actual” because ‘actual’ just means ‘thisworldly’, or ‘is part of my world’. For Lewis, ‘actual’ is an indexical term, like ‘I’ or ‘here’ or ‘now’. What ‘actual’ applies to on a given use depends on features of the context of utterance, in particular, on the speaker, and the speaker’s world. When I say of something that it is “actual,” I say simply that it is part of my world; Phillip Bricker 122 when my otherworldly counterpart says of something that it is “actual,” he says simply (if he is speaking English) that it is part of his world. I call my worldmates “actual” and my otherworldly counterparts “merely possible”; my counterpart calls his world- mates “actual” and me “merely possible.” And we all speak truly, just as many people in different locations all speak truly when each says, “I am here.” For Lewis, being actual or merely possible does not mark any ontological distinction between me and my counterparts because – as with being here or being there – there is no ontological distinction to be marked. What sort of property is expressed by a given use of ‘actual’ on Lewis’s account? When Lewis’s indexical analysis of actuality is combined with his analysis of world, we get that, in any context, ‘actual’ expresses the property of being spatiotemporally related to the speaker in that context. Thus, in any context, the property expressed by ‘actual’ is a relative property, a property things have in virtue of their relations to things, not in virtue of how they are in themselves. That makes actuality, on Lewis’s account, doubly relative: what property is expressed by a given use of ‘actual’ is rela- tive to the speaker; and the property thus expressed is itself a relative property. Lewis’s indexical account confl icts rather severely with our ordinary way of think- ing about actuality. As Robert Adams vividly put it: “We do not think the difference in respect of actuality between Henry Kissinger and the Wizard of Oz is just a differ- ence in their relations to us” (Adams 1974: 215). According to Lewis, however, a believer in concrete worlds has no choice but to accept an indexical analysis of actu- ality according to which actuality is doubly relative. For, Lewis argues, if my use of ‘actual’ instead expressed an absolute property that I have and my otherworldly counterparts lack, then no account could be given of how I know that I am actual. I have counterparts in other worlds that are epistemically situated exactly as I am; whatever evidence I have for believing that I have the supposed absolute property of actuality, they have exactly similar evidence for believing that they have the property. But if no evidence distinguishes my predicament from theirs, then I don’t really know that I am not in their predicament: for all I know, I am a merely possible person falsely believing myself to be absolutely actual. Thus, Lewis concludes, accepting concrete worlds together with absolute actuality leads to skepticism about whether one is actual. Since such skepticism is absurd, a believer in concrete worlds should reject absolute actuality.28 Lewis’s indexical account makes short work of the skeptical problem, and that is an argument in its favor. On Lewis’s analysis, ‘I am actual’ is a trivial analytic truth analogous to ‘I am here’. Just as it makes no sense for me to wonder whether I am here (because ‘here’ just means ‘the place I am at’), so it makes no sense for me to wonder whether I am actual (because ‘actual’ just means, according to Lewis, ‘part of the world I am part of’). Moreover, my otherworldly counterparts have no more trouble knowing whether they are actual than I do. When one of my counterparts says, “I am actual,” he speaks truly (if he is speaking English), and he knows this simply in virtue of knowing that he is part of the world he is part of. Thus, Lewis can explain why it strikes us as absurd for someone to wonder whether or not she is actual. Is Lewis correct, however, that a believer in concrete worlds has no choice but to reject absolute actuality? I hope not. I, for one, could not endorse the thesis of a plu- rality of concrete worlds if I did not hold that there was a fundamental ontological Concrete Possible Worlds 123 distinction between the actual and the merely possible. A Lewisian approach to modal- ity that rejects absolute actuality does not seem to me to be tenable. Actuality, I claim, is a categorial notion: whatever belongs to the same fundamental ontological category as something actual is itself actual. When Lewis insists, then, that all worlds are ontologically on a par, this can only be understood as saying that all worlds are equally actual – his denials notwithstanding. But that undercuts Lewis’s defense of concrete worlds: an analysis of modality as quantifi cation over concrete parts of actuality, no matter how extensive actuality may be, is surely mistaken.29 A way to test whether actuality is absolute or merely relative is to ask whether it is coherent to suppose that actuality is composed of island universes: parts that stand in no spatiotemporal (or other external) relations to one another. If actuality is abso- lute, the hypothesis of island universes is coherent: something could be actual even though entirely disconnected from the part of actuality we inhabit. But if actuality is merely relative, as Lewis supposes, then the hypothesis of island universes is analyti- cally false – and that seems wrong. Nor would it help for Lewis to switch to the amended analysis of possibility suggested in section 2: that analysis would make the hypothesis of island universes metaphysically possible – true at some plurality of worlds; but the hypothesis would remain analytically false of (what Lewis calls) actu- ality. Accepting that combination compounds the problem, rather than solving it. Fortunately, I think a Lewisian can accept absolute actuality without falling victim to the skeptical problem. To see how, we fi rst need to distinguish, for any predicate, the concept associated with the predicate from the property expressed by a given use of that predicate. The concept associated with a predicate is naturally identifi ed with its meaning. It embodies a rule that determines, for each context of use, what property is expressed by the predicate in that context. A predicate is indexical if it expresses different properties relative to different contexts of use; in that case, the associated concept can also be called indexical. Indexicality is one kind of relativity: relativity to features of context. But that sort of relativity must be distinguished from the relativity of the property expressed. Lewis’s argument against absolute actuality presupposes that these two sorts of relativity must go together, that any indexical analysis of actuality will be doubly relative. But that assumption, I think, is mistaken. Indexical concepts can be associated with predicates that express either relative or absolute properties. For example, consider the indexical predicate ‘is a neighbor’. On different occasions of use, it expresses different properties. When I use the predicate, it expresses the property of being one of my neighbors; when you use the predicate, it expresses the property of being one of your neighbors. On any use, the property expressed is a relative property: whether a person has the property expressed depends on that per- son’s relations to the speaker. Other indexical predicates, however, express absolute properties on each occasion of use. For example, the indexical predicate ‘is nutritious’ expresses different properties relative to different speakers (depending on age, or state of health). But, on each use, the property expressed is absolute, not relative: something is nutritious (for the speaker) in virtue of its chemical nature, not in virtue of its rela- tive properties; if two things are chemical duplicates of one another, then either both or neither are nutritious (for the speaker). Now, on Lewis’s analysis of actuality, the concept associated with the predicate ‘is actual’ is indexical, and the property expressed by the predicate, on each context of Phillip Bricker 124 use, is relative. It is the indexicality of the concept that allows for a solution to the skeptical problem. It is the relativity of the property that leads, I have claimed, to an untenable position. Is there a way of analyzing actuality so that the concept is indexi- cal but the property is absolute? Consider this: ‘is actual’ in my mouth expresses the property, “belonging to the same fundamental ontological category as me.” That builds the categorial nature of actuality directly into the analysis. But, thanks to the indexi- cal component, it makes short work of the skeptical problem. On this analysis, I know that I am actual simply in virtue of knowing that I belong to the same ontological category as myself. Knowledge that I am actual is just as trivial as on Lewis’s analysis of actuality, as it should be, but the property of actuality remains ontologically robust.30 Lewis would no doubt object that a theory of concrete worlds with absolute actual- ity is less parsimonious than his own. Granted. What matters, however, is which theory gets it right. If actuality is a categorial notion, as I believe, then Lewis’s indexical theory must be rejected. Lewis would also, I suspect, object that the notion of absolute actuality is mysterious. Perhaps. (It is not, however, the mystery of primitive modality: absolute actuality is no more primitive modality than is absolute truth.) It does not help our understanding, for example, to say that merely possible entities are “less real” than actual entities: both merely possible and actual entities exist, and I do not understand how existence could be a matter of degree. Nor does it help to say that actual and merely possible entities exist in different ways, that there are two modes of existence: if that means anything at all, it just means that there are two funda- mental ontological categories. The best inroad to understanding the distinction between the actual and the merely possible, I would say, comes from considering how we and our surroundings differ from what exists merely as an object of our thought. Lewis himself allows that there are entities of distinct ontological categories: indi- viduals and sets. The distinction between an individual and its singleton is arguably no less mysterious than the distinction between an actual thing and its merely possible qualitative duplicates. In the case of sets, Lewis embraces the mystery.31 Why not also, in the case of possibilia, embrace the mystery of absolute actuality? The answer turns on whether concrete talking donkeys and fl ying pigs are any easier to believe in if they belong to a different ontological category than actual donkeys and pigs. I leave that to the reader to ponder.
5 Modality De Re and Counterparts
It is traditional to divide modal statements into two sorts: de dicto and de re. A modal statement is de dicto, it is sometimes said, if the modal operator applies to a proposi- tion (Latin: dictum); it is de re if the modal operator applies to a property to form a modal property, which is then attributed to some thing (Latin: res). Thus, ‘necessarily, all birds are feathered’ is de dicto; ‘Polly is necessarily feathered’ is de re. The tradi- tional characterization, however, is defective in a number of ways. For one thing, equivalent statements are not always classifi ed alike. Indeed, any de dicto modal statement is equivalent to a statement attributing a modal property to the (actual) world. Thus, ‘necessarily, all birds are feathered’ is equivalent to ‘the world is Concrete Possible Worlds 125 necessarily such that all birds are feathered’, which the above criterion classifi es as de re. Moreover, a de re modal statement such as, ‘Polly is necessarily feathered’ is equivalent to ‘necessarily, Polly is feathered’ (or, perhaps, ‘necessarily, if Polly exists, Polly is feathered’), which the above criterion classifi es as de dicto. Another defect of the traditional characterization is that it fails to provide an exhaustive classifi cation of modal statements: modal statements with a complicated structure will not be clas- sifi ed either as de dicto or de re. A better way to characterize the de dicto/de re distinction looks to the content of modal statements, rather than their form. Here is the rough idea, neutrally expressed so as to apply to Lewisians and non-Lewisians alike. All possible worlds theorists will have to provide an account of how truth at a world is to be determined, how a world represents that things are one way or another. Part of any such account will involve providing for each world a domain of entities – the entities that in some primary sense “inhabit” the world – and saying, for each entity in a world’s domain, what properties it has at the world. Any such account will also have to say how it is deter- mined, when an entity is picked out as belonging to the domain of one world, whether the entity exists at some other world, and what properties it has at this other world. Call this “crossworld representation de re.” Now, what makes a modal statement de re is that in the course of evaluating its truth or falsity, one must have recourse to facts about crossworld representation de re. A modal statement is de dicto, on the other hand, if no recourse to crossworld representation de re is needed to evaluate its truth or falsity. To illustrate with a standard example: compare the de re ‘everything is necessarily material’ with the de dicto ‘necessarily, everything is material’. The former statement depends on crossworld representation de re: it says that every entity in the domain of the actual world is material, not only at the actual world, but at every possible world (better: at every possible world at which it exists). The latter statement does not depend on crossworld representation de re: it says that at every possible world, everything in the domain of that world is material. These two state- ments are not equivalent. The de dicto statement is made false by a possible world whose domain contains a non-material object; but if that possible world doesn’t represent de re concerning any actual object that it is non-material, then the de re statement may still be true. How is crossworld representation de re determined? The simplest answer, of course, would be this: an entity picked out as belonging to the domain of one world exists at some other world just in case it also belongs to the domain of that other world. On this account, the domains of different worlds overlap, and all facts about what properties an entity has at a world are given directly by how that world represents that entity to be. To exist at a world is just to belong to the domain of that world. I will say, following standard though somewhat misleading usage, that such an account endorses “transworld identity.” To illustrate, consider George W. Bush. On the tran- sworld identity theory, Bush belongs not only to the actual domain, but also to the domain of many merely possible worlds. Some of these worlds represent him as having properties he doesn’t actually have, such as losing the presidential election in 2004, or being a plumber. The Lewisian, however, needs a different account of crossworld representation de re. For the Lewisian, each world has as its domain just the entities that are part of Phillip Bricker 126 the world. Bush is part of the actual world, and is in the actual domain. But since worlds do not overlap, Bush is not in the domain of any merely possible world. How, then, does a merely possible world represent Bush as existing and having properties he doesn’t actually have? Lewis responds: by having in its domain a counterpart of Bush. A merely possible world represents de re concerning Bush that he exists and, say, is a plumber by containing a counterpart of Bush that is a plumber. So, for the Lewisian, existing at a world must be distinguished from being in the domain of a world: Bush exists at many worlds, although he is in the domain of – is part of – only one. Because Bush exists at many worlds, the Lewisian can be said to accept “transworld identity” in a weak, uncontroversial sense; but not in the stronger sense that requires overlapping domains. What makes an entity in one world a counterpart of an entity in another? Accord- ing to Lewis, the counterpart relation is a relation of qualitative similarity. He writes:
Something has for counterparts at a given world those things existing there that resemble it closely enough in important respects of intrinsic quality and extrinsic relations, and that resemble it no less closely than do other things existing there. Ordinarily something will have one counterpart or none at a world, but ties in similarity may give it multiple counterparts. (Lewis 1973: 39)
With a counterpart relation in place, de re modal statements can be analyzed in terms of concrete worlds and their parts. For example, to consider the simplest cases: ‘Bush is possibly a plumber’ is true just in case at some world some counterpart of Bush is a plumber; ‘Bush is necessarily (or essentially) human’ is true just in case at every world every counterpart of Bush is human.32 For the Lewisian, we have a simple way of distinguishing de re from de dicto: de re modal statements depend for their evalu- ation on the counterpart relation; de dicto modal statements do not. Whether one is a Lewisian or not, there are good reasons to prefer counterpart relations to transworld identity as an account of representation de re. I have space here to consider just one such reason, namely, that only counterpart relations can allow for essences that are moderately, without being excessively, tolerant.33 Let me explain. Modality de re is the realm of essence and accident. A property had by an individual is essential to the individual if that individual couldn’t exist without the property; it is accidental if it is not essential. An individual’s essence is the sum of all its essential properties. These notions will be translated into the framework of the counterpart theorist and the transworld identity theorist in different ways. For the counterpart theorist, an essential property of an individual is a property had by the individual and all of its counterparts. For the transworld identity theorist, an essential property of an individual is a property had by the individual itself at every world at which it exists. These two accounts come apart if the counterpart relation does not have the logical properties of identity, such as transitivity.34 When they come apart, the counterpart theorist enjoys a fl exibility that the transworld identity theorist cannot match. In particular, only a counterpart theorist can allow that an individual could have been somewhat different, say, with respect to its material composition or its origins, but could not have been wildly different. The transworld identity theorist will have to hold that essential properties are either not tolerant at all (with respect Concrete Possible Worlds 127 to material composition, or origins) or excessively tolerant; moderation will have to be abandoned. And that, in many cases, will lead to the wrong truth conditions for de re modal statements. To illustrate the problem of moderately tolerant essences, consider the following simple, though somewhat implausible, example. Suppose that it is essential to any person to have at least one of the (biological) parents he or she in fact has, but that it is not essential to have both. Thus, I could have had a different mother, and I could have had a different father, but I couldn’t have had both a different mother and a different father. My essence, then, is moderately tolerant with respect to my origins. Now, if representation de re works by transworld identity, an essence such as this leads to contradiction. Call my mother m and my father f. Since my essence is (mod- erately) tolerant, I could have had a different father and the same mother. So there is a world w and a person p existing at w such that p = me and p has father f ′ (≠ f ) and p has mother m. But now, since p’s essence is (moderately) tolerant, p could have had a different mother and the same father. So there is a world w′ and a person p′ existing at w′ such that p′ = p and p′ has mother m′ (≠ m) and p′ has father f ′. But if p′ = p and p = me, then p′ = me (by the transitivity of identity). So, at w′ I exist and my father is f ′ and my mother is m′. I could have had both a different father and a different mother after all, which contradicts the supposition that my essence is moderately tolerant. What to do? It would not be plausible, I think, to deny that individual essences can be moderately tolerant. A better solution is to switch to counterpart theory. If repre- sentation de re works by counterpart relations instead of transworld identity, then moderately tolerant essences are unassailable. A counterpart relation is based (at least in part) on qualitative similarity, and relations of similarity are not in general transitive. For the example at hand, the counterpart theorist will simply say that although p is a counterpart of me and p′ is a counterpart of p, p′ is not a counterpart of me. Thus, the counterpart theorist is not driven to assert that I could have had both a different mother and a different father, and no contradiction can be derived.35 So much in favor of counterpart theory; now for something on the other side. Many philosophers have argued, the theoretical benefi ts of counterpart theory not- withstanding, that counterpart theory provides unacceptable truth conditions for de re modal statements. If Lewisian realism is committed to counterpart theory, they say, so much the worse for Lewisian realism. I will briefl y consider two of the most common lines of attack, both of which can be traced to Saul Kripke’s infl uential dis- cussion of counterpart theory in Naming and Necessity.36 For the fi rst line of attack, consider Kripke’s complaint that according to counterpart theory:
[I]f we say “Humphrey might have won the election (if only he had done such-and-such),” we are not talking about something that might have happened to Humphrey, but to someone else, a ‘counterpart’. Probably, however, Humphrey could not care less whether someone else, no matter how much resembling him, would have been victorious in another possible world. (Kripke 1980: 45)
Kripke’s objection naturally falls into two parts. The fi rst part is that, on the analysis of modality de re provided by counterpart theory, the modal property, might have won the election, is attributed to Humphrey’s counterpart rather than to Humphrey himself. Phillip Bricker 128 But surely, the objection continues, when we say that “Humphrey might have won,” we mean to say something about Humphrey. This part of the objection, however, is easily answered. According to counterpart theory, Humphrey himself has the modal property, might have won the election, in virtue of his counterpart having the (non- modal) property, won the election. Moreover, that Humphrey has a winning counterpart is a matter of the qualitative character of Humphrey and his surroundings; so on the counterpart theoretic analysis, the modal statement is indeed a claim about Humphrey. The second part of Kripke’s objection is more troublesome. We have a strong intu- ition, not only that the modal statement, “Humphrey might have won the election,” is about Humphrey, but that it is only about Humphrey (and his surroundings). On coun- terpart theory, however, the modal statement is also about a merely possible person in some merely possible world; and that, Kripke might say, is simply not what we take the modal statement to mean. The fi rst thing to say in response is that the charge of unintuitiveness would apply equally to any theory that uses abstract worlds to provide truth conditions for modal statements; for our intuitive understanding of modal state- ments such as “Humphrey might have won the election” does not seem to invoke abstract worlds any more than counterparts of Humphrey. The objection, then, if it is good, would seem to cut equally against all possible worlds approaches to modality; if anything, it would favor a non-realist approach that rejects possible worlds, concrete or abstract. But is the objection good? Should our pre-theoretic intuitions as to what our statements are and are not about carry much, or even any, weight? I think not. A philosophical analysis of our ordinary modal statements must assign the right truth values and validate the right inferences; moreover, it must be able to withstand mature philosophical refl ection. But requiring that philosophical analyses match all our pre- theoretic intuitions would make systematic philosophy all but impossible. The second line of attack on counterpart theory I want to consider also has its origin in Kripke’s Naming and Necessity; but I will present the argument in the way I fi nd most effective, even though it may not coincide with Kripke’s intentions. The argument begins with the observation that we often simply stipulate that we are considering a possibility for some given actual individual. For example, we can simply say: consider a possible world at which Bush lost the election in 2004. In doing so, we consider a possible world that represents de re concerning Bush that he lost. Of course, such stipu- lation cannot run afoul of Bush’s essential properties: we cannot stipulate that a world represents de re concerning Bush that he is a poached egg if Bush is essentially human. The point, rather, is that such stipulation may be legitimate even if no loser at the world in question can be singled out as qualitatively most similar to Bush. But then, the argument concludes, representation de re cannot be based (entirely) on relations of qualitative similarity. If counterpart relations are relations of qualitative similarity, as Lewis asserts, then counterpart theory must be rejected. To illustrate the argument, consider the possibility that I have an identical twin. It seems coherent to suppose that, in the possibility being considered, neither I nor my twin is qualitatively more similar to the way I actually am than is the other. Nonethe- less, in the possibility being considered, I am one of the twins and not the other; indeed, we can stipulate that I am the fi rst-born twin. How can the Lewisian account for such a possibility? It seems that the Lewisian has to hold that there is a possible world that represents de re concerning me that I am the fi rst-born twin without Concrete Possible Worlds 129 representing de re concerning me that I am the second-born twin. But if representa- tion de re works by counterpart relations, that would seem to be impossible. Both twins are equally good candidates to be my counterpart, if the counterpart relation is a relation of qualitative similarity. A Lewisian has the following perfectly adequate response. If counterpart relations are relations of qualitative similarity, then indeed each twin is a counterpart of me at the world in question. Given that, the world can only represent de re concerning me that I am the fi rst-born twin if it also represents de re concerning me that I am the second-born twin, lest representation de re not be determined by counterpart relations. But the Lewisian can allow that the one world represents two distinct pos- sibilities for me: it represents de re concerning me that I am the fi rst born of two identical twins in virtue of containing a counterpart of me that is a fi rst-born identical twin; but it also represents de re concerning me that I am the second born of two identical twins in virtue of containing a counterpart of me that is the second born. In this way, all the facts of crossworld representation de re still depend only on the one qualitative counterpart relation; but when there are multiple counterparts at a world, multiple possibilities are represented within a single world.37 There is another sort of example involving stipulation de re, however, that the above response does nothing to accommodate. Not only can we stipulate that we are considering a possibility that involves a given individual, we can also stipulate that we are considering a possibility that does not involve a given individual. And, with this sort of stipulation, there do not seem to be any qualitative constraints. Indeed, it is perfectly legitimate to say: consider a possibility qualitatively indiscernible from actuality but in which I do not exist. In the possibility envisaged, I have a doppel- ganger, a person exactly like me in every qualitative respect, intrinsic and extrinsic; but that person isn’t me. I fi nd this intuition compelling, and think that any account of modality de re must fi nd a way to accommodate it. But now the counterpart theorist is in trouble if counterpart relations are relations of qualitative similarity. If the pos- sibility in question is represented by some non-actual world qualitatively indiscernible from the actual world, then an inhabitant of that non-actual world is qualitatively indiscernible from me without being my counterpart. If the possibility in question is instead somehow represented by the actual world, then there would have to be a counterpart relation under which I am not a counterpart of myself. Either way, the Lewisian would have to reject the idea that counterpart relations are (always) relations of qualitative similarity. So be it. That is a retreat from Lewis’s original understanding of counterpart theory, but it is by no means a defeat for the Lewisian. Counterpart theory, fi rst and foremost, is a semantic theory for providing truth conditions for de re modal statements. As such, it should adapt to those de re modal statements we take to be true. As long as this is accomplished in a way that doesn’t compromise the metaphysics of Lewisian realism, nothing of value is lost.38
6 Conclusion
When Lewis fi rst began advocating the thesis that there exists a plurality of concrete worlds, he received in response mostly “incredulous stares.” That soon changed. Over Phillip Bricker 130 the ensuing years, arguments for and against Lewisian realism have fi lled philosophi- cal books and journals. Lewisians have had to develop and revise their position in the light of powerful criticism; non-Lewisian alternatives have sprouted like weeds in the philosophical landscape. The debate goes on; as with other metaphysical debates, a decisive outcome is not to be expected. And through it all, the incredulous stares remain: Lewisian realism does disagree sharply, as Lewis himself concedes (Lewis 1986: 133), with common-sense opinion as to what there is. There seems to be a fundamental rift – unbridgeable by argument – between ontologically conservative philosophers who have, what Bertrand Russell called, “a robust sense of reality,” and ontologically liberal philosophers who respond, echoing Hamlet: “there is more on heaven and earth than is dreamt of in your philosophy.” No doubt, the Lewisian approach to modality will always be a minority view. But the power and elegance of the Lewisian approach has been widely appreciated by philosophers of all stripes. The bar is set high for the assessment of alternative views.
Notes
1 The fullest statement of Lewis’s theory of possible worlds is contained in his magnum opus, On the Plurality of Worlds (1986). Lewis’s view is sometimes called “modal realism.” 2 Historically, it was the attempt to provide semantics for modal logic that catapulted pos- sible worlds to the forefront of analytic philosophy. The locus classicus is Kripke (1963). 3 The locus classicus of the eliminative approach to modality is “Reference and Modality” in Quine (1953). 4 Of course, there is much more to modality than statements of (metaphysical) possibility and necessity. But the project of paraphrasing more complex modal locutions in terms of possible worlds has also met with considerable success. 5 For a discussion of the four ways, see Lewis (1986: 81–6). 6 An object is fully determinate if and only if, for any property, either the property or its negation holds of the object. In the case of worlds, this is equivalent to: for any proposi- tion, either the proposition or its negation is true at the world. 7 Because logical relations between propositions can be represented by relations between classes of worlds; for example, one proposition logically implies another just in case the class of worlds at which the one is true is included in the class of worlds at which the other is true. But see note 18. 8 The analysis of modality de re in terms of counterpart relations was fi rst introduced in Lewis (1968). 9 What it means for the concept of actuality to be indexical, and how that relates to whether the property of actuality is relative or absolute, is discussed in section 4. 10 See Lewis’s discussion of “magical ersatzism” (1986: 174–82). 11 I use ‘ideology’ roughly in Quine’s sense; see Quine (1953: 130–2). But I do not suppose our total theory is couched in fi rst-order predicate logic. Thus, all primitive terms of the language contribute to the ideology, not just the primitive predicates. 12 The perfectly natural properties make for intrinsic qualitative character; the perfectly natural relations are the fundamental ties that bind together the parts of worlds. (See also note 20.) For discussion of the notion of a perfectly natural property, and a defense of its legitimacy, see Lewis (1986: 59–69). 13 Lewis (1986: 75–6) sketches such an analysis. (He calls them the “analogical spatiotemporal relations,” but I drop the ‘analogical’.) Concrete Possible Worlds 131 14 For (a version of) the argument that Lewisian analyses of modality are circular, see McGinn (2000: 69–74). For a response, see Bricker (2004). 15 Lewis’s brief argument for this occurs in a footnote in Lewis (1986: 7). For further discus- sion, see Stalnaker (1996). Of course, the argument presupposes classical logic, that our total theory satisfi es the law of non-contradiction. For a non-classical approach to con- tradictions, see Priest (1998). 16 For example, according to one such principle, for any individual, simple or composite, it is possible that a duplicate of that individual exist all by itself. Applying this principle to a disconnected sum leads to the possibility of island universes. See Bricker (2001: 35–7) for detailed argument. Lewis (1986) does not explicitly accept any principle this strong; but Langton and Lewis (1998: 341) accept such a principle, and claim that it is part of the combinatorial theory put forth in Lewis (1986). 17 This view is presented and defended in Bricker (2001). Note that, on this approach, if island universes actually exist, then there is more than one actual world. But I will continue to speak of the actual world for ease of presentation. 18 Note that on a view that accepts mathematical systems alongside the concrete worlds, it is natural to distinguish between logical, metaphysical, and mathematical modality. Logical modality is absolute modality: it quantifi es over concrete worlds and abstract systems both. Metaphysical and mathematical modality are both restricted modality, quantifying over just the concrete worlds or just the mathematical systems, respectively. The term ‘logical space’ is now best used to refer to mathematical and modal reality together, not just modal reality. 19 I do not accept that there are any entities that are abstract in virtue of violating (C4). There is a mental operation of abstraction, which involves ignoring some features and attending to others; and we can represent the results of this mental procedure, if we want, by using set-theoretic constructions in ways made familiar by mathematicians. But, in my view, there are no “indeterminate objects.” 20 A property is qualitative (in the narrow sense) if its instantiation does not depend on the existence of any particular object, but does depend on the instantiation of some perfectly natural property. (The broad sense drops the second clause.) Every object a instantiates the non-qualitative property: being identical with a. Every natural number other than zero instantiates the non-qualitative (structural) property: being the successor of something. I suppose that the perfectly natural properties are all qualitative, and that the qualitative properties supervene on the perfectly natural properties and relations. All qualitative properties are categorical, but not conversely. 21 This “simple fi x” presupposes that material objects can be identifi ed with regions of space- time. That is, it presupposes that worlds do not divide into two distinct domains: an immaterial space-time, and material objects that occupy regions of space-time. On the “dualist” view, the immaterial space-time regions, arguably, would not instantiate any perfectly natural properties. 22 Or, in Quine’s slogan: “To be is to be the value of a variable.” That is to say, we are onto- logically committed to those entities that belong to the domain over which the variables of our quantifi ers range. See Quine (1953: 13). 23 Lewis (1986: 5–69) surveys some of the uses to which possible worlds have been put in systematic philosophy. Lewis’s oeuvre taken altogether provides a monumental testament to the fruitfulness of possible worlds. 24 Chapter 2 of Lewis (1986) argues that the cost of accepting concrete worlds is manageable by responding to eight objections from the literature; chapter 3 argues that rival views that take worlds to be abstract entities all have serious objections. For summaries of some of these arguments, see Bricker (2006a). Phillip Bricker 132 25 For an introduction to truthmaking, see Armstrong (2004). For reasons to accept only a weak truthmaker principle that applies to positive truths, see Lewis (2001). 26 For an introduction to the logical and metaphysical issues raised by this notion of inten- tionality, see Priest (2005). 27 We can call on the method of supervaluations to explain why I can truly say that there is one thing that I am thinking about: a dodecahedron made of solid gold. See Lewis (1993). 28 Lewis (1970) fi rst introduced his indexical theory of actuality and invoked the skeptical argument to support it. See also Lewis (1986: 92–6). 29 As Lewis himself concedes: “if the other worlds would be just parts of actuality, modal realism [Lewis’s brand of realism about possible worlds] is kaput” (1986: 112). 30 For a detailed attempt to develop an alternative indexical analysis of actuality on which actuality is absolute, see Bricker (2006b). 31 See Lewis (1991), especially pp. 29–35 on “mysterious singletons.” 32 Not all de re modal attribution follows this pattern. For example, ‘Bush necessarily exists’ should be analyzed as the falsehood, ‘at every world there is some counterpart of Bush’, not the trivial truth, ‘at every world every counterpart of Bush exists’. For discussion, see Lewis (1986: 8–13). 33 Two other important arguments supporting counterpart theory are the following. (1) Con- tingent Identity Statements. Only counterpart theory allows one to hold, for example, that a statue is identical with the lump of clay from which it is made, even though one can truly say: “that statue might have existed and not been identical with that lump of clay.” (2) Inconstancy of Representation De Re. Only counterpart theory allows one to hold, in accordance with ordinary practice, that attributions of essential properties may vary from context to context. In both of these cases, the counterpart theorist introduces multiple counterpart relations to achieve the desired result. See Lewis (1986: 248–59). 34 A relation is transitive if and only if whenever one thing bears the relation to a second, and the second bears the relation to a third, then the fi rst bears the relation to the third. 35 The problem of moderately tolerant essences was fi rst introduced in Chisholm (1967). The counterpart-theoretic solution is discussed in Lewis (1986: 227–35). 36 See Kripke (1980: 44–53). For other well-known objections to counterpart theory, see Plantinga (1973). 37 Lewis (1986: 243–8) presents and defends this response. 38 Anti-Haecceitism – the view that representation de re supervenes on qualitative features of worlds (in the broad sense) – is arguably an essential component of the metaphysics of Lewisian realism. But the Lewisian need not abandon anti-Haecceitism to accommodate the possibility that things could be qualitatively the same as they actually are and yet nothing actual exist. The Lewisian can say that the actual world represents this possibility with respect to a counterpart relation under which nothing is a counterpart of anything. Although, contra Lewis, this counterpart relation is not a relation of qualitative similarity, it is nonetheless qualitative (in the broad sense): it does not distinguish between qualita- tive indiscernibles. See Lewis (1986: 220–35) for his characterization and defense of anti-Haecceitism.
References
Adams, Robert. 1974. “Theories of Actuality,” Nous 8: 211–31. Repr. in Loux (1979). Armstrong, D. M. 2004. Truth and Truthmakers (Cambridge: Cambridge University Press). Bricker, Phillip. 1996. “Isolation and Unifi cation: The Realist Analysis of Possible Worlds,” Philosophical Studies 84: 225–38. Concrete Possible Worlds 133 ——. 2001. “Island Universes and the Analysis of Modality,” in Gerhard Preyer and Frank Siebelt, eds., Reality and Humean Supervenience: Essays on the Philosophy of David Lewis (Lanham, MD: Rowman and Littlefi eld). ——. 2004. “McGinn on Non-Existent Objects and Reducing Modality,” Philosophical Studies 118: 439–51. ——. 2006a. “David Lewis: On the Plurality of Worlds,” in John Shand, ed., Central Works of Philosophy. Volume 5: The Twentieth Century: Quine and After (Chesham: Acumen). ——. 2006b. “Absolute Actuality and the Plurality of Worlds,” in John Hawthorne, ed., Philo- sophical Perspectives 2006, Metaphysics (Oxford: Blackwell). Chisholm, Roderick. 1967. “Identity Through Possible Worlds: Some Questions,” Nous 1: 1–8. Repr. in Loux (1979). Kripke, Saul. 1963. “Semantical Considerations on Modal Logic,” Acta Philosophical Fennica 16: 83–93. ——. 1980. Naming and Necessity (Cambridge, MA: Harvard University Press). Langton, Rae, and Lewis, David. 1998. “Defi ning ‘Intrinsic’,” Philosophy and Phenomenological Research 58: 333–45. Lewis, David. 1968. “Counterpart Theory and Quantifi ed Modal Logic,” Journal of Philosophy 65: 113–26. Repr. in Loux (1979). ——. 1970. “Anselm and Actuality,” Nous 4: 175–88. ——. 1973. Counterfactuals (Oxford: Basil Blackwell). ——. 1986. On the Plurality of Worlds (Oxford: Basil Blackwell). ——. 1991. Parts of Classes (Oxford: Basil Blackwell). ——. 1993. “Many, But Almost One,” in Keith Campbell, John Bacon, and Lloyd Reinhardt, eds., Ontology, Causality, and Mind: Essays on the Philosophy of D. M. Armstrong (Cambridge: Cambridge University Press). ——. 2001. “Truthmaking and Difference-Making,” Nous 35: 602–15. Loux, Michael. ed. 1979. The Possible and the Actual: Readings in the Metaphysics of Modality (Ithaca, NY: Cornell University Press). McGinn, Colin. 2000. Logical Properties: Identity, Existence, Predication, Necessity, Truth (Oxford: Oxford University Press). Plantinga, Alvin. 1973. “Transworld Identity or Worldbound Individuals?” in Milton Munitz, ed., Logic and Ontology (New York: New York University Press). Repr. in Loux (1979). Priest, Graham. 1998. “What is So Bad About Contradictions?” Journal of Philosophy 95: 410–26. ——. 2005. Towards Non-Being: The Logic and Metaphysics of Intentionality (Oxford: Oxford University Press). Quine, Willard Van Orman. 1953. From a Logical Point of View (Cambridge, MA: Harvard University Press). Stalnaker, Robert. 1976. “Possible Worlds,” Nous 10: 65–75. Repr. in Loux (1979). ——. 1996. “Impossibilities,” Philosophical Topics 24: 193–204.
Phillip Bricker 134 CHAPTER 3.2
Ersatz Possible Worlds
Joseph Melia
1 Introduction
The philosophical benefi ts that possible worlds offer are rich indeed. Everyday modal statements such as ‘there are many different ways the world could have been’ can be taken at face value, as talking about possible worlds other than the actual one. A wide range of modal concepts can be analyzed in terms of possible worlds in a logic that is familiar and well understood. Problematic and capricious de re modal state- ments can be tamed and understood. Physical necessity, deontic obligation, and other strengths of modality can be given unifying analyses in terms of possible worlds. Previously unanswerable questions about modal validity can be resolved. Once obscure intensional logics can be given a possible worlds model theory, and completeness and soundness results will have genuine philosophical signifi cance.1 Unifying ontological reductions – propositions as sets of possible worlds, properties as sets of possible individuals – become available once we help ourselves to possible worlds. Though not conclusive, the fact that possible worlds enable us to unify and simplify our theories in such ways speaks in their favor. True, one can wonder whether the simplifi cation and systematization that result are reason for thinking the resultant theory more likely to be true; harshly stated, such criteria can appear to be aesthetic rather than rationally compelling. But appeals to simplicity and unifi cation are not restricted to philosophical theories; they appear in the theoretical sciences and in certain parts of common sense. When one doubts whether such theoretical virtues are worth having, one runs the risk of thereby being skeptical about a great deal more than just possible worlds. For all this, it is hard to accept David Lewis’s view of possible worlds. On Lewis’s view, merely possible worlds are like the actual one, concrete island universes con- taining – well, just about anything you care to think of, really. If it’s possible, it’ll literally be part of one of Lewis’s possible worlds. The view that there is an infi nite number of concrete island universes containing talking donkeys and stalking centaurs sits uneasily with common sense. The view that every possible object exists is an appalling violation. Though the kinds of simplicity and unifi cation that possible worlds bring to our theories may be theoretical virtues worth having, simplicity of ontology, of the entities that a theory postulates, is a theoretical virtue too. Here, Lewis’s theory of possible worlds scores very badly.2 One could grant the entire case for possible worlds’ theoretical utility, grant that the theoretical benefi ts are great, yet still rationally believe that they are simply not worth the massive ontological costs. If only there were a way of getting all, or most, of the theoretical benefi ts that possible worlds have to offer without the excessive ontological costs and appalling violation of our common-sense beliefs, the case for possible worlds might be restored. Enter the ersatzer. Like Lewis, the ersatzer is a realist about possible worlds: possible worlds exist, can be referred to and quantifi ed over in our theories and analyses. But the ersatz possible worlds have quite a different nature from Lewis’s possible worlds. There may be no ontological free lunch – perhaps the ersatzer will have to invoke unreduced propositions, or states of affairs, or properties – but the ersatzer hopes his theory will be ontologically a lot cheaper than Lewis’s and a whole load easier to believe in. The ersatzer’s theory may not yield all the benefi ts that Lewis’s theory offers (primitive modal concepts, for example, are hard to eliminate altogether on the ersatzer’s scheme) but the ersatzer’s laudable aim is to construct an ontologically parsimonious theory of possible worlds capable of getting as many of the theoretical benefi ts as possible. It is essentially this goal, rather than any particular thesis about the nature of possible worlds, that unifi es the ersatzers. We cannot characterize the ersatzer as one who believes that worlds are abstract rather than concrete, as there are versions of ersatzism where possible worlds come out concrete. We cannot characterize the ersatzer as one who rejects mere possibilia – things that don’t actually exist but that could have – for there’s no reason why the ersatzer couldn’t have ersatz possibilia along with his ersatz possible worlds. Ersatzism is better seen as a program rather than a particular unifi ed position in the philosophy of possible worlds, and there are a number of different versions on the market.
2 The Ersatzer’s Zoo
Even if modality cannot be analyzed non-circularly, all ersatzers agree on the following: